{"id":40096,"date":"2026-05-07T18:18:18","date_gmt":"2026-05-07T22:18:18","guid":{"rendered":"https:\/\/studentwork.prattsi.org\/infovis\/?p=40096"},"modified":"2026-05-07T20:03:36","modified_gmt":"2026-05-08T00:03:36","slug":"supply-chain-mapping","status":"publish","type":"post","link":"https:\/\/studentwork.prattsi.org\/infovis\/visualization\/supply-chain-mapping\/","title":{"rendered":"Supply Chain Mapping"},"content":{"rendered":"\n
\"\"<\/figure>\n\n\n\n
<\/div>\n\n\n\n

Goals<\/strong><\/h3>\n\n\n\n

My goal for this visualization was to use Gephi to visualize a supply chain for a particular greenhouse producer, Mastronardi\/Sunset Produce, and to show relationships between production, distribution and retail nodes. My longer term goal is to map where there is leverage to support organizing by greenhouse workers. On a personal level, I wanted to work with a smaller dataset to get more comfortable with Gephi, which I find more challenging than the other tools used so far in class.<\/p>\n\n\n\n

I was inspired by the visualizations of supply chains uploaded to the Manifest: Supply Chain Platform<\/a>. However, I wanted to be careful not to show a relationship which may not yet be confirmed through research (i.e. showing an unconfirmed relationship between a particular distribution hub and a retailer).<\/p>\n\n\n\n

<\/div>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
<\/div>\n\n\n\n

Dataset<\/strong><\/strong><\/h3>\n\n\n\n

I wanted to try a data scraping approach to create my own dataset for this visualization. Through previous corporate research, I had created a list of all greenhouses and warehouses for this employer using job ads and industry press. Along with other colleagues, I did store research to track where the Sunset products were sold and generated a list of major retailers across North America. I was curious what using AI to do data scraping would generate in comparison. <\/p>\n\n\n\n

I asked ChatGPT to create a list of retailers selling Sunset cucumbers as well as Sunset greenhouses and warehouses facilities and populate the data in a spreadsheet that could be uploaded to Gephi. As I am still getting my head around how to set up an Edge and Node table for this type of relationship, this generated spreadsheet was helpful to build my understanding of the backend structure.<\/p>\n\n\n\n

First iteration<\/strong><\/strong><\/strong><\/h3>\n\n\n\n

I uploaded the initial spreadsheet to see what visuals would be generated. While I tried to use the ForceAtlas 2 layout, I could not get the correct scale to see the network clearly underneath all the node labels. I found the Fruchterman Reingold layout easier to work with and to create a visualization that allowed me to see the overall structure of how the relationships had been set up. 
<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

I realized that the node table generated by ChatGPT was creating different nodes for different types of cucumber products, and although this could be interesting for a subsequent visualization, I did not need that level of granularity and it made it difficult to see the overall supply chain. I ran the request again but with the instruction to condense multiple cucumber product nodes into one type of cucumber.<\/p>\n\n\n\n

Second iteration<\/strong><\/strong><\/strong><\/strong><\/h3>\n\n\n\n

For my second version, I colored the nodes by type to more clearly show the different elements of the supply chain (ie. Production, Distribution and Retail).  Because I don\u2019t have details on production or sales volumes, I didn\u2019t think it was appropriate to have nodes or edges sized differently. A future version could integrate that data to show more important nodes in the supply chain, or larger flows of products. Similarly, I didn’t have the data to look for clusters or other patterns.<\/p>\n\n\n\n

<\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>\n\n\n\n

This version helped me to see the overall structure, but also where there were some missing edges, or mis-categorized relationships. I could also see where ChatGPT had missed data I had collected manually. Having the corporation as one node, and the product as a second node helped keep the production, distribution, and retail elements of the supply chain distinct. However, I wanted to see what would happen if the product was eliminated as a node and the corporation was the primary source node.\u00a0<\/p>\n\n\n\n

I generated this alternate spreadsheet through ChatGPT, but I found the resulting visualization too simplistic, and not as useful for showing the supply chain flow. I tried multiple variations, but ultimately decided to stick with the second iteration, and to manually add missing information and correct the edges.<\/p>\n\n\n\n

Other examples<\/strong><\/strong><\/strong><\/strong><\/strong><\/strong><\/h3>\n\n\n\n

Interestingly, as I looked for other examples of supply chain visualizations created using Gephi at this point in my design process, I found examples where researchers had similarly used a product as a central node or hub (such as in this mapping of a cadmium telluride solar cell supply chain<\/a>), or a geographic location (such as this H&M garment supply chain map<\/a>).\u00a0 Both visualizations used a legend to help decipher the different types of nodes, and I realized this would improve the legibility of my visualization.<\/p>\n\n\n\n

<\/div>\n\n\n
\n
\"\"<\/figure>\n<\/div>\n\n\n
<\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>\n\n\n\n

Third iteration<\/strong><\/strong><\/strong><\/strong><\/h3>\n\n\n\n

In my final version I manually added additional nodes in Gephi generated through my own research and the corresponding edges. I also cleaned up missing or inconsistent edges, and deleted some redundant nodes and incorrect edges. I refined the color palette to ensure each broad category was grouped by color. I exported a SVG file to be able to add a legend in Illustrator and do further design refinements.<\/p>\n\n\n\n

<\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>\n\n\n\n

Reflection \/ Next steps<\/strong><\/h3>\n\n\n\n

I now have a better understanding of how to structure an edge and node table, and therefore a stronger grounding to imagine future visualizations of supply chains. I realize that first sketching out on paper various ways to represent the data would have been a better starting point, rather than trying to work it out through different iterations in Gephi.\u00a0<\/p>\n\n\n\n

Overall, I feel my network visualization broadly met my goal of representing my available information on the Sunset supply chain. I can see how patterns of possible leverage emerge when visualizing the data this way, for example, the concentration of retailers selling Sunset products under various retail parent companies. On the other hand, the visualization doesn\u2019t show any other broad patterns due to the lack of quantitative data attached to the nodes or edges. The nodes also feel very condensed, and their placement somewhat random.<\/p>\n\n\n\n

I would like to experiment more with the ForceAtlas 2 layout and the scale and gravity options to create more white space. In my current iteration, I felt that the black background made the colors stand out, but it also impacted legibility of the node labels. In a future iteration, I would like to try using a white background (I didn\u2019t know how to edit this without altering other elements!).<\/p>\n\n\n\n

I would also like to show volume or concentration and more clustering of the nodes. In the absence of reliable product volume or sales data, I could try layering in data on greenhouse size and size the nodes accordingly. I could cluster nodes by geography and type (for example for greenhouses where I do have a specific location information).\u00a0 Placing the network visualization on top of a map would be my ultimate goal, knowing that this requires longer term data collection on the direct connections between nodes as well as their geographic location.<\/p>\n\n\n\n

Sources<\/h6>\n\n\n\n

MacCarthy, B. L., Ahmed, W. A. H., & Demirel, G. (2022). Mapping the supply chain: Why, what and how? International Journal of Production Economics, 250<\/em>, 108688.<\/a>
Nuss, P., Graedel, T. E., Alonso, E., & Carroll, A. (2016). Mapping supply chain risk by network analysis of product platforms<\/em>. Sustainable Materials and Technologies, 10<\/em>, 14\u201322.<\/a><\/p>\n\n\n\n

<\/div>\n","protected":false},"excerpt":{"rendered":"

Goals My goal for this visualization was to use Gephi to visualize a supply chain for a particular greenhouse producer, Mastronardi\/Sunset Produce, and to show relationships between production, distribution and retail nodes. My longer term goal is to map where there is leverage to support organizing by greenhouse workers. On a personal level, I wanted…<\/p>\n","protected":false},"author":5055,"featured_media":40097,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[149,342,1],"tags":[],"coauthors":[1921],"class_list":["post-40096","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-labs","category-networks","category-visualization"],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/studentwork.prattsi.org\/infovis\/wp-content\/uploads\/sites\/3\/2026\/05\/sunset-v7-scaled.png?fit=2560%2C2222&ssl=1","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/paBdcV-aqI","_links":{"self":[{"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/posts\/40096","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/users\/5055"}],"replies":[{"embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/comments?post=40096"}],"version-history":[{"count":16,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/posts\/40096\/revisions"}],"predecessor-version":[{"id":40223,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/posts\/40096\/revisions\/40223"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/media\/40097"}],"wp:attachment":[{"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/media?parent=40096"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/categories?post=40096"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/tags?post=40096"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/studentwork.prattsi.org\/infovis\/wp-json\/wp\/v2\/coauthors?post=40096"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}