data presentation techniques geography a level

Percentage or Divided Bar Charts

Stacked Bar Charts

Min-Max Range Floating Bar Charts

Beach profiles (using slope angle data).

Beach Profiles (using height change data)

Beach Profiles (using absolute height data)

BiHistograms

Bi-Polar Charts

Box Plots or box and whisker charts (Horizontal)

Box plots (vertical).

Box Plots (Outliers)

Box Plots (Mean and Standard Deviation)

Cumulative Frequency Chart

Donut (Doughnut) Charts

Kite diagrams (any data entities).

Kite Diagrams (Species Abundance)

Line charts.

Map Cross Sections

Polar Area Charts

Polar Charts

Polar Chart Overlays

Polar Scatter Charts

Proportional circles chart.

Rose Diagrams

Sand Dune Profiles (using slope angle data)

Slope Profiles (using slope angle data)

Word Clouds

• Geographical skills and enquiry
• Human geography
• Physical geography
• Careers Spotlights
• Skills Boosts

Webinars //

Skills boost, presenting fieldwork data, hosted by chloe searl.

In this skills boost, Chloe looks at what we mean by 'appropriate' data presentation and how to optimise this to boost your fieldwork grade. She looks at how to present spatial and qualitative data, and ways of combining these methods to create standout graphics that grab people's attention and most importantly showcase the geography.

Presented by Chloe Searl, the Island Geographer and writer of the three accompanying texts on Methods of Collecting, Analysing and Presenting Fieldwork Data. Find them in the GA Shop here .

Share with friends

Get in Touch

The Geographical Association 160 Solly Street, Sheffield, S1 4BF

0114 296 0088 [email protected]

Stay Social

Cookies Privacy Terms Sitemap

© 2020 The Geographical Association. Charity no. 1135148. Company no. 07139068. Website by QWeb Ltd

Become a member and discover where geography can take you.

• Resources for ...
• A student guid...

A student guide to the A Level independent investigation (Non-examined Assessment - NEA)

The following documents are available in the Downloads section below:

Download a copy of the guide below

Lawrlwythwch gopi o'r canllaw

Before you start

i   – Independent Investigation - Student Planning Form        Independent Investigation - Student Planning Form (Word document) ii  – A Guide to Writing A Research Plan iii – A Guide to Effective Background Reading iv – A Guide to Referencing

Section 1 – Introduction

1a – A Guide to Hypotheses

Section 2 – Data collection

2a – A Guide to Different Types of Data

2b – A Guide to Data Collection Techniques – Quantitative Measuring

2c – A Guide to Data Collection Techniques – Surveys

2d – A Guide to Data Collection Techniques – Interviews

2e – A Guide to Data Collection Techniques – Questionnaires

2f – A Guide to Data Collection Techniques – Observations, Photographs and Field Sketches

2g – A Guide to Sampling Techniques

2h – A Guide to Recording Data in the Field

2i – A Guide to Avoiding Biased Data

Section 3 – Data presentation

3a – A Guide to Column Charts and Histograms

3b – A Guide to Pictograms

3c – A Guide to Pie Charts

3d – A Guide to Scatter and Line Graphs

3e – A Guide to Box and Whisker Graphs

3f – A Guide to Kite Diagrams

3g – A Guide to Triangular graphs

3h – A Guide to Rose and Radial Graphs

3i – A Guide to Isoline Maps

3j – A Guide to Choropleth Maps

3k – A Guide to Pictorial Data Presentation

3l – A Guide to Presenting Qualitative Data

Section 4 – Data analysis

4a – A Guide to Measures of Central Tendency

4b – A Guide to Measuring Proportions

4c – A Guide to Measures of Dispersion

4d – A Guide to Cost-Benefit Analysis

4e – A Guide to Spearman’s Rank

4f – A Guide to Chi-Squared Testing

4g – A Guide to Simpson’s Diversity Index

4h – A Guide to Pearson’s Product Moment

4i – A Guide to Nearest Neighbour Analysis

4j – A Guide to Mann-Whitney U Test

4k – A Guide to Qualitative Data Analysis

Section 5 – Conclusions

Section 6 – Evaluation

Section 7 – Final checks 

Posters – A guide to reading your research landscape

A copy of FAQs from the awarding bodies

View a video from the Royal Met Soc - MetLink - Weather Fieldwork for your A Level Geography Independent Investigation

File name Files

Full Guide: Guide to the NEA

Student Planning Form - Word

Student Planning Form - PDF

A Guide to Writing A Research Plan

A Guide to Effective Background Reading

A Guide to Referencing

1a – A Guide to Hypotheses

2a – A Guide to Different Types of Data

2b – A Guide to Data Collection Techniques – Quantitative Measuring

2c – A Guide to Data Collection Techniques – Surveys

2d – A Guide to Data Collection Techniques – Interviews

2e – A Guide to Data Collection Techniques – Questionnaires

2f – A Guide to Data Collection Techniques – Observations, Photographs and Field Sketches

2g – A Guide to Sampling Techniques

2h – A Guide to Recording Data in the Field

2i – A Guide to Avoiding Biased Data

3a – A Guide to Column Charts and Histograms

3b – A Guide to Pictograms

3c – A Guide to Pie Charts

3d – A Guide to Scatter and Line Graphs

3e – A Guide to Box and Whisker Graphs

3f – A Guide to Kite Diagrams

3g – A Guide to Triangular graphs

3h – A Guide to Rose and Radial Graphs

3i – A Guide to Isoline Maps

3j – A Guide to Choropleth Maps

3k – A Guide to Pictorial Data Presentation

3l – A Guide to Presenting Qualitative Data

4a – A Guide to Measures of Central Tendency

4b – A Guide to Measuring Proportions

4c – A Guide to Measures of Dispersion

4d – A Guide to Cost-Benefit Analysis

4e – A Guide to Spearman’s Rank

4f – A Guide to Chi-Squared Testing

4g – A Guide to Simpson’s Diversity Index

4h – A Guide to Pearson’s Product Moment

4i – A Guide to Nearest Neighbour Analysis

4j – A Guide to Mann-Whitney U Test

4k – A Guide to Qualitative Data Analysis

Section 7 – Final checks

Posters – A guide to reading your research landscape

Poster – RMetS weather and climate

FAQs from the awarding bodies

A Guide to Collecting Weather Data

BSG geomorphology projects for students

NEA word limit letter

Ethical Data resources

Doug Specht and colleagues from the University of Westminster have produced resources to provide some extra guidance and support around ethics and data use for the NEA

This project was funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily those of the Foundation

Back To Top

• 0 Shopping Cart

Data Presentation in Geography

Data Presentation Techniques – How and Why?

Divided Bar Chart

Pyramid Pyramid

Compound Line Graph

Scatter Graphs

Dispersion Graphs

Choropleth Maps

Coming Soon – Data Presentation Techniques – How and Why?

Located Bar Chart

Desire Line

Proportional Symbols

Coming soon! – I want to…

Share this:.

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)
• Click to share on Pinterest (Opens in new window)
• Click to email a link to a friend (Opens in new window)
• Click to share on WhatsApp (Opens in new window)
• Click to print (Opens in new window)

Please Support Internet Geography

If you've found the resources on this site useful please consider making a secure donation via PayPal to support the development of the site. The site is self-funded and your support is really appreciated.

Search Internet Geography

Latest Blog Entries

Pin It on Pinterest

• Click to share
• Print Friendly

How to Get an A* in an A-Level Geography NEA

In A-Level by Think Student Editor May 22, 2023 Leave a Comment

An NEA is worth 20% of your overall grade in A-Level Geography. It may not seem like much at first, but that’s 1/5 of your qualification – so you definitely want to make sure it boosts your overall grade! As well as this, there will be hundreds of other students writing A-Level Geography coursework, so yours needs to stand out amongst the others. But, if you were like me, this might be the first time you’ve ever had to complete coursework for an A-Level. So how do you do well?

In this article, I’ll be taking you through (as a former A-Level Geography student) all the dos and don’ts for your geography NEA, and the advice you need to get an A*!

Table of Contents

What is an A-Level Geography NEA?

The geography NEA is the mandatory coursework, that is a part of A-Level Geography. All UK exam boards require A-Level Geography students to produce an NEA. It is also referred to as an “Independent Investigation”.

For an NEA, you will choose a question related to physical or human geography, and then you will collect data to help you answer this question . Most NEAs are around 3,000- 4,000 words. They are essentially research papers!

For inspiration make sure you check out this Think Student article with 75+ NEA ideas!

How is an A-Level Geography NEA structured?

At the front of your geography NEA, you will have to attach a cover sheet provided by your exam board. This will usually have your name, candidate number, centre number, and your title question on it. It must be signed by you and your teachers.

An A-Level Geography NEA typically has around 7 sections :

• Introduction to the investigation
• Methodology/data collection
• Data presentation, analysis, interpretation and evaluation
• Evaluation of the investigation
• Bibliography

Exam boards are not too strict on formatting, however, so as long as you have all the major areas covered, you can format these how you like.

Your bibliography should contain all the references for any secondary material you used as part of your NEA. When you submit an NEA, you will be asked to declare that the work you have produced is 100% your own, and your NEA will be checked for plagiarism . Make sure to reference properly!

As for the rest of the NEA sections, I’ll be explaining them in more detail later in the article, so don’t worry!

How do you get an A* in an A-Level Geography NEA?

Getting an A* in your A-Level Geography NEA is not easy, but it’s definitely not impossible.

The UK exam boards will have their NEA criteria up on their website. NEA marking criteria is usually broken down into 4 “levels” (with level 1 being the least marks and level 4 being the most marks), for each section of the NEA.

To achieve an A* in an A-Level Geography NEA, you’ll need to meet the level 4 criteria in most if not all sections of your NEA.

As a former A-Level Geography student, below I’ll share with you my advice on how to achieve the top marks in your NEA.

The A-Level Geography NEA introduction

The introduction to your A-Level Geography NEA is one of the most important parts – it sets up the rest of your investigation and shows examiners why they should keep reading!

Your introduction will outline your argument and will vaguely demonstrate what you are going to say and why this is important. Remember, you don’t want to say too much, because you’ve got the rest of the NEA to write!

Your introduction should also link to your title question; how is your argument going to relate to and answer your question?

The key to a good geography NEA introduction is to be concise and keep it simple. You should ask yourself: ‘if someone who doesn’t do A-Level Geography read my introduction, would they know what I’m going to talk about?’

How to write an A* A-Level Geography NEA introduction

Your introduction should break down your purpose for the investigation. It isn’t like your typical essay introductions which are 100-200 words – your NEA introduction should be up to 500 words.

It could be helpful to break down your title question into three or four “key inquiry questions”, which you can answer throughout your NEA.

You should also explain your title question, why you chose it and how your research is useful in thinking about the future of the research your question tackles.

Your introduction is the opportunity to provide the examiner with details about your location; you could use maps (as these count as a form of data presentation), point out key geographical features, etc. You should give some local (relevant to your area) context and global context for the issue your question is answering.

To round off the introduction, include some basic geographical theory. For example, if your NEA investigation focuses on erosion, explain the different theories of erosion and how these apply to your investigation. This is an important demonstration of knowledge!

Remember, you can format your NEA however you like (within reason), so you can put this information in whichever order you like. Just make sure you cover all the key areas of your investigation!

The A-Level Geography NEA methodology

Your NEA methodology is a breakdown of how you collected the data you use and present in your coursework.

Your methodology will be one of the most detailed parts of your NEA. This may be surprising, but it’s because your methodology is used to show that your data is legitimate and collected properly.

A methodology is included in the majority of research papers, and your A-Level Geography NEA is no exception! Make sure you put time and care into writing your methodology properly, or it could undermine your investigation.

How do you write an A* methodology for an A-Level Geography NEA?

The way you physically present your methodology is up to you, but it should cover all the qualitative data (non-measurable data), and quantitative data (measurable/numerical data).

For example, I presented my methodology as a big table across 2 pages of my NEA. Definitely don’t underestimate the size of your methodology – it’s what verifies that your data is legitimate!

In your methodology, you should include:

• The types of data you collected
• Where you collected this data (collection points)
• The equipment you used to collect your data
• A description of the method
• How often you collected data (intervals)
• The sampling technique (stratified, systematic, etc.)
• A justification for your method

As part of your methodology, you should also include what are called “ethical considerations” and a “risk assessment”.

Ethical considerations essentially means showing awareness of any ethical problems with your data collection methods. As an example, if you used a survey as a data collection, a problem with that may be that the participants’ privacy is not protected. Therefore, an ethical consideration would be to anonymise the survey.

A risk assessment is an awareness of the risks that are involved with data collection (such as getting lost, injured, weather events, etc.), and what you will do to prevent these risks. For example, having an emergency contact.

Data presentation, analysis, interpretation and evaluation in an A-Level Geography NEA

The data section of your NEA is the longest chunk and is worth the most marks. Now that you’ve set up your investigation, this is the section where you present all of your findings and interpret them, by explaining what they show and why.

Don’t panic if not all of the data you collected can be used – I certainly had a bunch of random data I didn’t need by the end! Try to use as much data as possible, and different types of data.

This section helps establish your argument; it’s essentially the evidence for your conclusion as well as just being the body of your NEA.

Since this is a long section, it’s helpful for you and your examiner to split it up into chunks using subheadings. It’s not a good idea to signpost, for example putting the subheading “Analysis”. Instead, you might divide up your data by the location, or the method you used to collect it.

How do you present data to get an A* in an A-Level Geography NEA?

Data presentation in a geography NEA is probably the most unique part of the process – you get to present your data however you want (in accordance with the exam board guidelines, of course)!

In the data presentation section, you need to display all the data you collected for your investigation. This can be in charts, graphs, tables, photos, and more.

The data needs to be readable, so your graphs should be labelled correctly, and your photos should have captions. If you’re using any data that isn’t yours, remember to reference it correctly.

It’s also a good chance to add a bit of colour, to make your A-Level Geography NEA look great!

How do you analyse and evaluate data to get an A* in an A-Level Geography NEA?

Your data analysis, interpretation, and evaluation section of your geography NEA is the most important section.

You should pick out key elements of the data and explain what they mean with regard to your NEA investigation question. How does the data you collected argue for/against your question?

Where applicable, it’s a good idea to calculate and explain medians, means, modes, and averages, to show that you aren’t just repeating what’s already in your presentation. You need to do something with the raw numbers, you definitely shouldn’t just relay your exact findings.

When you’re analysing, ask yourself the question: what does my data mean?

When you’re evaluating, ask yourself the question: how does my data answer my investigation?

By keeping these questions in mind when you’re interpreting your data, you can show the examiner that you can prove why your data is important and that you have a good understanding of analysis and evaluation.

Should you include statistical tests to get an A* in an A-Level Geography NEA?

The short answer to this question is: absolutely!

By now, you will have practiced a few statistical tests as part of the rest of A-Level Geography, such as Spearman’s Rank, the T-Test, Mann-Whitney U test, and standard deviation.

You should aim to use one or two stats tests when presenting the data, you collected for your geography NEA. There is no ‘right’ or ‘wrong’ stats test, so choose whichever is applicable for your data.

Statistical tests are a good demonstration of your analytical, interpretative and evaluative skills . By including a couple, you are showing the examiner that you have a clear knowledge of what the tests mean and why they’re useful!

If you struggle with the calculations, don’t be afraid to ask for help. Obviously other people can’t do it for you (remember that as part of submitting your geography NEA, you will have to testify that your work is entirely your own), but you can always ask to be shown how to do them!

Evaluating your A-Level Geography NEA investigation

Your A-Level Geography NEA investigation evaluation is slightly different to the evaluation of your data. In this section of your NEA, you should evaluate the success of the overall investigation.

You should discuss your locations and the methods you used to collect your data (both primary AND secondary data!). What was good about them? What wasn’t so good? If you had been somewhere else and used different methods, how might the outcome of your investigation have changed?

It’s also important to acknowledge the validity of your conclusions.

For example, you may not have had time or access to the correct resources to collect some really important data, that would’ve affected your outcome and potentially changed it. Showing an awareness of this helps build a more sophisticated and mature argument.

It’s important to note than an evaluation is not the same as a conclusion! You shouldn’t be summarising your research. Instead, explain the positives and negatives of your research choices.

The A-Level Geography NEA conclusion

Your conclusion is crucial because it ties together your methods, research, and analysis. Remember those “key inquiry questions” I mentioned earlier? Well now is the time to answer them!

Your NEA conclusion will answer your title question and provide the examiner with a neat, rounded summary of your investigation. By reading the conclusion, someone should be able to know the key parts of your argument and why they are important.

A conclusion is also a place to propose solutions – what can we do in future that we aren’t doing now? How might future events like climate change impact your research?

If there are relevant questions that could impact the outcome of your investigation, but you don’t have time to consider them in detail, put them in your conclusion. This shows the examiner that you have an awareness of micro- and macro-scale issues!

How do you write an A* A-Level Geography NEA conclusion?

Like most essay conclusions, your geography NEA conclusion will summarise your main arguments, what you found, and what your data means. It can also be a good place to ask any of the questions you still don’t have answers to.

You could start by going through your inquiry questions and writing “sub-conclusions” in response to them. Then, you should move on to the big conclusion: answering your title question.

In your conclusion, you should highlight the key things you found as a result of your research, broadly and specifically. Showing consideration to the “big” and “small” issues is good for showing your critical thinking skills!

Your conclusion should be about the same length as your introduction, give or take. If you start running out of things to say, don’t add things unnecessarily to fill the word count – your conclusion should be the most clear and concise part of your NEA.

Examples of A-Level Geography NEAs

Most, if not all, exam boards will have an “exemplar” coursework on their website. For example, I’ve linked the OCR exemplar coursework for you here , so check your exam board website for their exemplar!

The exemplar coursework is written and submitted by a real student, but it’s important to follow the mark scheme, not just copying someone else’s coursework. Your NEA will be checked for plagiarism!

Similarly, most schools keep exemplar coursework from each year, so if you need some inspiration, ask your teachers for the coursework they have.

If you’re struggling for ideas of what to write on for your Geography NEA, check out this Think Student article with 75+ ideas!

*To learn more about the A-Level Geography NEA, check out the specifications from the main exam boards, AQA , Pearson Edexcel and OCR by clicking on their respective links.

Using Geospatial Data ( AQA A Level Geography )

Revision note.

Value of Big Data

• Geospatial data is data information that has a location
• Much of this data come from big data sets
• The census is an example of a big data set as it surveys the entire country and needs computational analysis to start making sense of the patterns 
• The census asks a number of questions which creates lots of complexity in the data, another feature of big data sets 
• Understanding geospatial data and its different forms allows geographers to infer spatial patterns and see the relationship between people, environment and place 
•  A spatial pattern simply means that there is a pattern in the data based on the place 
• For example, many data indicators in the indicators reveal a north-south divide which is a spatial pattern
• Geospatial data can be  qualitative or quantitative  
• When analysing geospatial data it is important to compare one source with another to check for reliability 
• Maps are an excellent example of geospatial data but when using their consideration must be given to: 
• Whether the map is choropleth or proportional  
• How this will affect the visual representation of the spread

A choropleth map showing the proportion of people born in the UK from the 2021 census

• The choropleth map from the UK 2021 census shows the proportion of people born in the UK.  A map like this has abrupt boundaries that suggest a significant change as soon as a country boundary is crossed which is not likely to be the case in reality

Proportional Symbols map showing the results of the 2017 general election 

• The proportional symbols map plots the results of the 2017 general election, this type of geospatial data illustrates the difference between many places very well and shows data associated with a more specific location than a choropleth map can 
• However, the proportional symbols make it very difficult to calculate the actual value, even if there was a key, and the size of the symbols can obstruct the map underneath, making the positioning less accurate

When approaching your data analysis six markers you have to look to see if you find a relationship between the variable or figures you are shown.  It is not just about describing what you see in the figure but analysing if a relationship exists, if it is a strong relationship and if there are any outliers or anomalies to the relationship. 

Things to look for: 

• The general pattern, is a headline that could describe the figure in one statement 
• The most and least, are they the same in both figures? 
• If there is a relationship between the two figures is it a positive or negative one? 
• How strong the relationship is 
• The outliers or anomalies that do not fit the pattern or relationship

Quantitative Sources of Data

• Quantitative sources of data are numerical 
• They are objectively measured and can therefore be compared across space and often across time as well 
• Much of the quantitative data we have in the UK comes from local councils and the census  
• The census is a nationwide survey that is taken every ten years to collect information that creates a picture of all the households and people in England and Wales  
• Scotland has a separate census
• The first modern census was taken in 1841

• Another very popular quantitative data source to understand place characteristics is the Index of Multiple Deprivation (IMD)
• To create the IMD seven components of deprivation are considered and put together to create a single score of deprivation 
• These are: income, employment, education, health, crime, barriers to housing and services and living environment
• A composite measure like this captures a full range of variables that contribute to deprivation in an area and recognises that one measure is not enough to truly represent a place

the index of multiple deprivation for the Leeds area in 2019  

• It is easy to see which areas are in the most deprived 10% of the country and then in the other deciles
• IMD maps are choropleth maps using small areas called Lower Super Output Areas 

You've read 0 of your 0 free revision notes

Get unlimited access.

to absolutely everything:

• Downloadable PDFs
• Unlimited Revision Notes
• Topic Questions
• Past Papers
• Model Answers
• Videos (Maths and Science)

Join the 100,000 + Students that ❤️ Save My Exams

the (exam) results speak for themselves:

Did this page help you?

Author: Alex Lippa

Alex graduated from the University of Cambridge in 2013 with an MA in Geography. She took part in the TeachFirst teacher training programme and has worked in inner city London for her whole career. As a Head of Geography and has helped many students get through their exams. Not only has she helped students to pass but she has supported multiple students towards their own places at the University of Cambridge to study geography. Alex has also been a private tutor and written resources for online platforms during her career.

Skip to content

Get Revising

Join get revising, already a member.

data presentation

notes on different data types used in presenting data and factors worth considering before selecting a technique to present your data with

• Created by: Thegirlwhoknewtoomuch - Team GR
• Created on: 23-09-13 14:32
• Geographical skills and fieldwork

No comments have yet been made

Similar Geography resources:

Geography skills 3.0 / 5 based on 1 rating

Six Stages of a Geography Investigation 1.5 / 5 based on 3 ratings

OCR Geographical Skills 5.0 / 5 based on 15 ratings

Data presentation: types of data and factors worth considering 0.0 / 5

goegraphy skills paper 0.0 / 5

Geography 4A Skills 4.0 / 5 based on 1 rating

Data presentation: non spatial data 0.0 / 5

What type of data presentation should I use? 5.0 / 5 based on 3 ratings

data presentation- non spatial data 0.0 / 5

Small Scale Ecosystems 0.0 / 5

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Product demand prediction with spatial graph neural networks.

1. Introduction

2. related works, 2.1. e-commerce demand prediction, 2.1.1. characteristics of online advertising data, 2.1.2. demand prediction methods, 2.1.3. spatial statistics in demand prediction, 2.2. graph neural network in modeling spatial dynamics events, 3. methodology, 3.1. attributed graph prediction problem formulation, 3.2. constructing graph with spatial adjacency, 3.3. attention-aware message propagation, 3.4. node-level product demand prediction, 3.5. training objective, 4. experiments, 4.1. experiment setup, 4.1.1. datasets, overall introduction, detailed data information.

• item id: Id of a particular advertisement.
• user id: Id of a user.
• region: The region that Ads belong to.
• city: The city that a Ad belongs to.
• top-level category: The top-level ad category as classified by Avito’s ad model.
• fine-grain category: The fine-grain ad category as classified by Avito’s ad model.
• param 1: The first optional parameter from Avito’s ad model.
• param 2: The second optional parameter from Avito’s ad model.
• param 3: The third optional parameter from Avito’s ad model.
• title: The textual title for the Ad.
• description: The multi-sentence textual description for the Ad.
• price: The numerical value for the Ad’s price.
• item seq number: Ad sequential number for the user.
• activation date: The date that the Ad was placed onto the platform.
• user type: The type of the user, including Private, Company, and Shop.
• image: The Id code corresponding to the image which is tied to a jpg file in train jpg. Considering that not every Ad has an image, we don’t employ this feature for further analysis.
• image top 1: Avito’s classification code for the image.
• deal probability: The target variable. This is the likelihood that the ad will actually sell the item. It is not possible to verify every transaction with certainty, so the value of this column can be any floating point number from zero to one.

Data Analysis

• deal probability: An initial observation from the distribution histograms indicates a pronounced long-tail distribution issue with the deal probability feature. Notably, approximately 65 % of the ads exhibit a zero deal probability, signifying a substantial portion of ads that do not culminate in a transaction. Conversely, a minimal fraction of ads achieve a deal probability of 1, indicating a successful sale. This distribution suggests a high variance in the likelihood of deals being closed across the dataset.
• price: Prior to analysis, the price feature undergoes a logarithmic transformation to normalize its distribution. Post-transformation, the price distribution approximates a normal distribution, as evidenced by the histograms. This transformation mitigates the skewness originally present in the data, facilitating more meaningful statistical analysis and interpretation.
• region: The analysis of the region feature reveals a geographical disparity in ad postings. The Krasnodar region emerges as the most prominent area for ad postings, followed closely by the Sverdlovsk and Rostov regions. This distribution highlights the regional variances in marketplace activity, potentially influenced by factors such as population density and economic conditions.
• city: Delving into the city feature, we observe that the highest number of ads are posted in Krasnodar and Ekaterinburg cities. Subsequent rankings include Novosibirsk, Rostov-na-Donu, and Nizhny Novgorod cities. This urban-centric distribution underscores the role of major cities as hubs for online marketplace transactions, possibly attributed to their larger populations and higher internet penetration rates.
• top-level category: The top-level category feature analysis reveals a dominant preference for posting ads in the “Personal things” category, accounting for more than 0.6 million users. Following this, the categories for “Home and Cottages” and “Consumer Electronics” are notable, with approximately 0.2 million users posting ads in each. This distribution indicates a significant inclination towards selling personal belongings, with a notable interest in home-related items and electronics.
• fine-grain category: Within the subcategories, “Clothes, shoes, accessories”, “Children’s clothing and footwear”, and “Goods for children and toys” emerge as the top three, each with around 0.3 million postings. This detailed breakdown within the fine-grain category feature further elucidates consumer behavior, highlighting a strong market for personal and children-related items.

Click here to enlarge figure

• deal probability and region: Upon examining the relationship between deal probability and region , it is observed that the mean deal probability across all regions hovers around 15 % . This uniformity suggests that while regional factors may influence the volume of ads, they do not significantly differentiate the likelihood of a deal being closed. This finding could indicate that other factors beyond geographical location play a more pivotal role in influencing deal probability.
• deal probability and city: Similar to the observation with regions, the analysis of deal probability and city reveals that all cities also exhibit a mean deal probability of approximately 15 % . This consistency across cities further supports the notion that the likelihood of closing a deal is not heavily dependent on specific urban centers, highlighting the importance of looking beyond geographic specifics to understand deal closure dynamics.
• deal probability and top-level category: A more nuanced insight emerges from the relationship between deal probability and top-level category . The “Services” category stands out with the highest mean deal probability at 40 % , followed by “Transport” and “Animals” at 25 % . This distinction suggests that ads within the “Services” category are significantly more likely to result in a deal, possibly due to the inherent nature of services being in higher demand or more immediately consumable compared to physical goods. This disparity underscores the potential for tailoring strategies based on category-specific demand dynamics.
• deal probability and fine-grain category: The analysis of deal probability and user type reveals a notable difference in mean deal probabilities between Private Users ( 15 % ) and Shop Users ( 5 % ). This discrepancy suggests that ads posted by private individuals are three times more likely to close a deal than those posted by shops. This could be attributed to a variety of factors, including perceived trustworthiness, pricing differences, or the nature of the goods and services offered by these user types.

4.1.2. Baselines

• Generalized Linear Model(GLM) [ 18 ]: A foundational approach in statistical modeling, GLM extends traditional linear regression to support various types of distribution for the target variable, such as binomial and Poisson distributions. This model is pivotal for understanding the linear relationships between the features and the target demand, serving as a baseline to assess the incremental benefits of more complex models.
• XGBoost [ 19 ]: A highly efficient and scalable implementation of gradient boosting framework, XGBoost has gained popularity for its performance in various predictive modeling competitions. It leverages an ensemble of decision trees, optimized through gradient boosting, to capture non-linear relationships and interactions among features.
• LightGBM [ 20 ]: An advanced gradient boosting model that utilizes a novel tree-growing algorithm to enhance efficiency and scalability. LightGBM is designed to handle large-scale data, offering a faster training process without compromising on model accuracy.
• CatBoost [ 21 ]: Another gradient boosting variant, CatBoost is renowned for its handling of categorical features directly, without the need for extensive preprocessing. It provides robust solutions to avoid overfitting, making it highly effective in diverse predictive tasks, including demand prediction.
• Multiple Layer Perceptron (MLP) [ 22 ]: MLP is a class of feedforward artificial neural network (ANN) that consists of at least three layers of nodes: an input layer, a hidden layer, and an output layer. MLP utilizes a backpropagation technique for training, capable of capturing complex non-linear relationships between inputs and outputs.
• LSTM [ 23 ]: Long Short-Term Memory networks, a type of recurrent neural network (RNN) architecture, are specifically designed to address the vanishing gradient problem of traditional RNNs. LSTMs are adept at learning long-term dependencies, making them particularly suitable for time-series predicting tasks like demand prediction.
• GRU [ 24 ]: Gated Recurrent Units (GRUs) are a variant of RNNs that simplify the LSTM architecture while retaining its capability to capture dependencies over various time spans. GRUs offer a more efficient and equally effective alternative for sequential data modeling.
• CNN [ 25 ]: Convolutional Neural Networks, traditionally known for their prowess in image processing, have also been adapted for spatial predicting. By capturing spatial dependencies through their hierarchical structure, CNNs can be utilized to effectively model the geographical information and relationships in demand data.

4.1.3. Evaluation Metrics

• Mean Absolute Error (MAE): MAE is a straightforward metric that calculates the average absolute difference between the actual demand y i and the predicted demand y ^ i across all observations. It provides an intuitive measure of the model’s accuracy, with lower values indicating better performance. The MAE is defined as MAE = 1 N ∑ i = 1 N | y i − y ^ i | , (5) where N is the total number of observations. This metric is particularly useful for understanding the magnitude of prediction errors without considering their direction.
• Root Mean Squared Error (RMSE): This metric offers a more sensitive measure of model accuracy by squaring the errors before averaging, thus giving greater weight to larger errors. RMSE is defined as the square root of the average of squared differences between the predicted and actual values: RMSE = 1 N ∑ i = 1 N ( y i − y ^ i ) 2 , (6) where N represents the total number of observations. The RMSE is beneficial for identifying when a model might be prone to producing significant errors, as it penalizes larger discrepancies more heavily than smaller ones.
• R-squared (R 2 ): R-squared, also known as the coefficient of determination, measures the proportion of the variance in the dependent variable that is predictable from the independent variables. It indicates how well the regression predictions approximate the real data points. An R-squared of 1 indicates that the regression predictions perfectly fit the data. The formula for R-squared is R 2 = 1 − ∑ i = 1 N ( y i − y ^ i ) 2 ∑ i = 1 N ( y i − y ¯ ) 2 (7) where y ¯ is the mean of the actual demand values. R-squared is a relative measure of fit that can be useful when comparing different models’ abilities to explain the variability in the data.

4.1.4. Implementation Details

4.2. overall performance (rq1), 4.3. ablation study (rq2), 4.4. hyperparameter robustness (rq3), 5. conclusions and future works, 5.1. conclusions, 5.2. future works.

• Temporal Dynamics Integration: Incorporating temporal dynamics into the SGNN framework could significantly enhance the model’s predictive accuracy. Future work could explore methods for embedding time-series data into the graph structure, allowing the model to capture not only spatial relationships but also temporal patterns in consumer behavior and demand fluctuations.
• Hybrid Models: Combining SGNN with other machine learning techniques, such as reinforcement learning or unsupervised learning algorithms, could lead to hybrid models that leverage the strengths of multiple approaches. For instance, reinforcement learning could optimize inventory levels dynamically based on SGNN demand predictions, offering a comprehensive solution for supply chain management.
• Cross-Domain Adaptation: Exploring the applicability of SGNN in domains beyond retail, such as urban planning, transportation, and social network analysis, could unveil new insights and applications. The spatial and relational modeling capabilities of SGNN hold potential for predicting traffic flow, urban development trends, or information propagation in social networks.
• Advanced Graph Architectures: Investigating more sophisticated Graph Neural Network architectures, including Graph Attention Network and Heterogeneous Graph Neural Network, could provide deeper insights into complex spatial interactions. These advanced models could better capture the heterogeneity in data types and relationships present in retail networks.
• Scalability and Efficiency: Addressing the computational challenges associated with SGNN, particularly for large-scale applications, remains a critical area for future work. Developing more efficient algorithms and leveraging distributed computing frameworks could enhance the scalability and practicality of SGNN for real-world applications.
• Interpretability and Explainability: Enhancing the interpretability of SGNN models is crucial for gaining insights into the underlying factors driving demand predictions. Future work could focus on developing methodologies for visualizing and interpreting graph-based models, providing valuable feedback for decision-makers in retail and other sectors.

Author Contributions

Institutional review board statement, informed consent statement, data availability statement, conflicts of interest.

• Kim, J.U.; Kim, W.J.; Park, S.C. Consumer perceptions on web advertisements and motivation factors to purchase in the online shopping. Comput. Hum. Behav. 2010 , 26 , 1208–1222. [ Google Scholar ] [ CrossRef ]
• Rai, S.; Gupta, A.; Anand, A.; Trivedi, A.; Bhadauria, S. Demand prediction for e-commerce advertisements: A comparative study using state-of-the-art machine learning methods. In Proceedings of the 2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT), Kanpur, India, 6–8 July 2019; IEEE: Piscataway, NJ, USA. [ Google Scholar ]
• Kin-man To, C. Innovation process management in global fashion businesses: A review of contextual aspects. Res. J. Text. Appar. 2003 , 7 , 60–73. [ Google Scholar ] [ CrossRef ]
• Levis, A.A.; Papageorgiou, L.G. Customer demand forecasting via support vector regression analysis. Chem. Eng. Res. Des. 2005 , 83 , 1009–1018. [ Google Scholar ] [ CrossRef ]
• Jain, A.; Karthikeyan, V.; Sahana, B.; Shambhavi, B.R.; Sindhu, K.; Balaji, S. Demand forecasting for e-commerce platforms. In Proceedings of the 2020 IEEE International Conference for Innovation in Technology (INOCON), Bangluru, India, 6–8 November 2020; IEEE: Piscataway, NJ, USA, 2020. [ Google Scholar ]
• Tugay, R.; Oguducu, S.G. Demand prediction using machine learning methods and stacked generalization. arXiv 2020 , arXiv:2009.09756. [ Google Scholar ]
• Bandara, K.; Shi, P.; Bergmeir, C.; Hewamalage, H.; Tran, Q.; Seaman, B. Sales demand forecast in e-commerce using a long short-term memory neural network methodology. In Proceedings of the Neural Information Processing: 26th International Conference, ICONIP 2019, Sydney, NSW, Australia, 12–15 December 2019; Proceedings, Part III 26. Springer International Publishing: Berlin/Heidelberg, Germany, 2019. [ Google Scholar ]
• Zhang, G.; Patuwo, B.E.; Hu, M.Y. Forecasting with artificial neural networks: The state of the art. Int. J. Forecast. 1998 , 14 , 35–62. [ Google Scholar ] [ CrossRef ]
• Kuo, C.; Reitsch, A. Neural networks vs. conventional methods of forecasting. J. Bus. Forecast. 1995 , 14 , 17. [ Google Scholar ]
• Azzouni, A.; Pujolle, G. A long short-term memory recurrent neural network framework for network traffic matrix prediction. arXiv 2017 , arXiv:1705.05690. [ Google Scholar ]
• Aburto, L.; Weber, R. Improved supply chain management based on hybrid demand forecasts. Appl. Soft Comput. 2007 , 7 , 136–144. [ Google Scholar ] [ CrossRef ]
• Islek, I.; Ögüdücü, S.G. A Decision Support System for Demand Forecasting based on Classifier Ensemble. In FedCSIS (Communication Papers) ; FedCSIS: Sofia, Bulgaria, 2017. [ Google Scholar ]
• LeSage, J.; Pace, R.K. Introduction to Spatial Econometrics ; Chapman and Hall/CRC: Boca Raton, FL, USA, 2009. [ Google Scholar ]
• David, O. Geographically weighted regression: The analysis of spatially varying relationships. Geogr. Anal. 2003 , 35 , 272–275. [ Google Scholar ]
• Noel, C. The origins of kriging. Math. Geol. 1990 , 22 , 239–252. [ Google Scholar ]
• Wu, Z.; Pan, S.; Chen, F.; Long, G.; Zhang, C.; Philip, S.Y. A comprehensive survey on graph neural networks. IEEE Trans. Neural Netw. Learn. Syst. 2020 , 32 , 4–24. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Gandhi, A.; Aakanksha; Kaveri, S.; Chaoji, V. Spatio-temporal multi-graph networks for demand forecasting in online marketplaces. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases ; Springer International Publishing: Cham, Switzerland, 2021. [ Google Scholar ]
• Hastie, T.J.; Pregibon, D. Generalized linear models. In Statistical Models in S ; Routledge: London, UK, 2017; pp. 195–247. [ Google Scholar ]
• Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm Sigkdd International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 785–794. [ Google Scholar ]
• Ke, G.; Meng, Q.; Finley, T.; Wang, T.; Chen, W.; Ma, W.; Ye, Q.; Liu, T.Y. Lightgbm: A highly efficient gradient boosting decision tree. Adv. Neural Inf. Process. Syst. 2017 , 30 , 52. [ Google Scholar ]
• Prokhorenkova, L.; Gusev, G.; Vorobev, A.; Dorogush, A.V.; Gulin, A. CatBoost: Unbiased boosting with categorical features. Adv. Neural Inf. Process. Syst. 2018 , 31 . [ Google Scholar ]
• Amalnick, M.S.; Habibifar, N.; Hamid, M.; Bastan, M. An intelligent algorithm for final product demand forecasting in pharmaceutical units. Int. J. Syst. Assur. Eng. Manag. 2020 , 11 , 481–493. [ Google Scholar ] [ CrossRef ]
• Abbasimehr, H.; Shabani, M.; Yousefi, M. An optimized model using LSTM network for demand forecasting. Comput. Ind. Eng. 2020 , 143 , 106435. [ Google Scholar ] [ CrossRef ]
• Shu, W.; Zeng, F.; Ling, Z.; Liu, J.; Lu, T.; Chen, G. Resource demand prediction of cloud workloads using an attention-based GRU model. In Proceedings of the 2021 17th International Conference on Mobility, Sensing and Networking (MSN), Exeter, UK, 13–15 December 2021; IEEE: New York, NY, USA, 2021. [ Google Scholar ]
• Tang, Z.; Ge, Y. CNN model optimization and intelligent balance model for material demand forecast. Int. J. Syst. Assur. Eng. Manag. 2022 , 13 (Suppl. S3), 978–986. [ Google Scholar ] [ CrossRef ]

Share and Cite

Li, J.; Fan, L.; Wang, X.; Sun, T.; Zhou, M. Product Demand Prediction with Spatial Graph Neural Networks. Appl. Sci. 2024 , 14 , 6989. https://doi.org/10.3390/app14166989

Li J, Fan L, Wang X, Sun T, Zhou M. Product Demand Prediction with Spatial Graph Neural Networks. Applied Sciences . 2024; 14(16):6989. https://doi.org/10.3390/app14166989

Li, Jiale, Li Fan, Xuran Wang, Tiejiang Sun, and Mengjie Zhou. 2024. "Product Demand Prediction with Spatial Graph Neural Networks" Applied Sciences 14, no. 16: 6989. https://doi.org/10.3390/app14166989

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals