Literature Review

There has been quite a bit written over the past few years about the global semiconductor supply chain and its vulnerability. A lot of the attention around the subject started ramping up after COVID-19 and the U.S.-China trade war showed how reliant the world is on just a few countries for chip manufacturing.

Exiger’s 2025 article on semiconductors does a good job of laying out the big-picture risks from more of a supply chain industry perspective. It helps to explain why semiconductors are hard to make, and how concentrated production is in places like Taiwan and South Korea. It also discusses how these companies face threats such as cyberattacks, export and import restrictions, and climate-related disasters. It is not super data-heavy, but it helped me to better understand the types of real-world risks that are shaping how companies and countries produce chips.

Reports from the Semiconductor Industry Association (2021) lay out the risks of relying on such a globally distributed and interdependent supply chain. They make sure to point out that even a single point of failure, such as a factory shutdown or geopolitical disruption, could throw off production around the world. They follow up the Exiger article well by taking a more policy and structure focused approach. Their main point is that the U.S. and other regions need to invest more in domestic production to make the whole system more resilient.

Additionally, Deloitte’s 2025 Industry Outlook expands out on that by describing how fast demand is growing, particularly with the rise of AI and data centers. They also examine how that demand is pushing countries to invest more in their own chip production. The report focuses on economic trends and future growth, showing how countries are responding with new funding and policy incentives. This report contains a lot of forecast and big-picture economic trends, but it does not really get into the specific supply routes or trade flows.

After digging into these sources, I realized that while there is a lot of analysis that seems to have been done on risk and supply chain dynamics, what has not been examined as much is the actual trade flows. There is a clear understanding of the challenges, but not a lot of visual representation of the trade itself. My project aims to fill that gap by using real trade data and visual tools to give a clearer picture of how this global system actually works.

Data Collection

Once I had a clear picture of the risks and challenges from the different sources, I started collecting trade data to actually see how the semiconductor supply chain is moving across different countries and regions. The main goal of this part of the project was to take all of those high-level ideas and back them up with real numbers.

I pulled my main dataset from the UN Comtrade database. I pulled 2024 trade data using HS code 8541, which covers semiconductors like “Diodes, transistors and similar devices”. I filtered by exports and imports to get both sides of the trade flow and downloaded the datasets for the top 25 reporting countries by trade value. I used Excel and a little bit of Python to clean and organize the data, mostly just for formatting and creating a consistent layout to compare the data.

Taiwan’s data was unfortunately a bit more difficult to work with since it is not listed as a reporter on Comtrade. So after some searching I was able to pivot and use data from the Taiwan Ministry of Finance. Their database does not use HS codes in the same way, so instead I pulled export data from Category 16.1 “Parts of Electronic Products”. This is not a perfect match for HS code 8541, but based on what is included, it likely captures a lot of the same types of goods. After some thinking and feedback about whether to estimate the share that’s specifically semiconductors, I decided not to and I believe it is better to keep the data clean and not throw off the accuracy.

I also looked into some recent trade policy data to give the project just a little more context. I pulled from sources like the SIA and CSIS, which break down things like the CHIPS Act, export controls, and new manufacturing subsidies. I did not go super in depth into every policy, but I have highlighted a few of the bigger ones to give a better sense of how government decisions shape supply chains.

For visuals, I used Tableau to build different graphs comparing exports and imports across countries. I also constructed a web map using Google CoLab where I show trade flows between key countries. This involved some interesting work to figure out which map tiles and formatting make the data easiest to understand. Also, I built out a full Tableau dashboard that ties the visuals together so people can actually explore the data themselves.

Overall, the methods I employed are mostly quantitative, but I tried to make sure all the data is readable and ties back into the bigger conversation. My main goal was to take abstractions about semiconductor policy and supply chain risks and actually visualize it using real data.