Close
Dillion Lee is studying a BA Geography with Social Data Science at UCL. Dillion completed a UCL Social Data Institute internship with the Climate Policy Initiative in 2024 and details his experience.
Describe the day-to-day responsibilities you had during your internship.
I was initially tasked with sourcing data to enrich an existing set of development finance datasets and to 'top-up' data that was missing from these datasets. I was also tasked with constructing a dataset of climate budgets. After this, I was tasked with coding a machine learning algorithm using a Lasso-Logistic Regression to tag climate finance investments in Python. These investments were then cleaned and assembled into a much larger African Climate Finance dataset that would then be analysed.
I was also involved in writing several reports on carbon markets, climate budgets and the costs of inaction on climate change, which were done concurrently with the data processing.
What project(s) were you involved in? What outcomes or deliverables were generated e.g. reports, articles, insight generated for the rest of the organisation etc.?
Assembling a dataset of climate investments, writing reports on carbon markets, climate budgets and costs of inaction on climate change. These reports and the dataset will be used in the Landscape of Climate Finance in Africa 2024 (AF24) report.
What was your favourite task/responsibility during your internship? Which piece of work are you most proud of?
I enjoyed learning python and using it to process and assemble a comprehensive dataset of climate investments. Being able to code in a new language, seeing the algorithms run and the final dataset process was incredibly satisfying. Applying the knowledge I gained in POLS0010 Data Analysis (Logistic Regression and Lasso Regression) to this task was also interesting as while datasets given to us for coursework and tutorials are relatively clean and can be processed immediately, this task required a complete end-to-end development of the algorithm, from building the training database to running and testing the algorithm, to fine-tunning specific parameters of the model.
I was also proud of the carbon market report I wrote, which is going to be published in the AF24 report at the end of the year.
What did you find challenging during your internship?
Handling large datasets and manual imputation of large amounts of data from various sources to enrich the datasets was tedious and felt daunting at times. Standardising the formatting of data from different sources was also complex.
What software and data analysis techniques did you have the opportunity to use?
Software: Python, RStudio and Microsoft Excel
Data Analysis: Data wrangling (data joins, handling missing data etc.) and machine learning (Lasso and Ridge Regression, Logistic Regression)
What was it like working for your provider?
Very challenging yet enjoyable. I never felt like I was being pressured to do more than I could and the working hours were incredibly flexible, with hybrid and staggered working arrangements. The office was also really nice and had good amenities. There was also a team day where we took part in a scavenger hunt around London. It gave me a chance to meet and bond with other members of the team and to talk to them and understand what some of the other workstreams CPI undertakes.