Automating Redseer's Monthly Data Workflow
Compressed several days of repetitive data work into a reusable firm-wide tool, freeing analyst time for actual insight generation
Pranjal Gujral
Analyst at Redseer Strategy Consultants
From their time as
Analyst
Redseer Strategy Consultants • 2026
Overview
Pranjal noticed early in his time at Redseer that the first seven working days of every month were consumed by the same set of tasks: cleaning and structuring a large data dump received from the tech team, which was used to triangulate historical numbers and validate hypotheses. The work was repetitive, the datasets were large enough to cause his laptop to lag, and none of it required judgment. It was pure grunt work sitting in front of the actual analysis.
The Story
Pranjal noticed early in his time at Redseer that the first seven working days of every month were consumed by the same set of tasks: cleaning and structuring a large data dump received from the tech team, which was used to triangulate historical numbers and validate hypotheses. The work was repetitive, the datasets were large enough to cause his laptop to lag, and none of it required judgment. It was pure grunt work sitting in front of the actual analysis.
He decided to automate it. Using Claude to help write the Python and HTML scripts, he built an initial cleaning script that handled the most time-consuming part of the process. He showed it to his manager, who was impressed and encouraged him to keep going. He then automated a second repetitive task, then a third, ending up with three to four Python scripts that handled the bulk of the monthly data preparation.
Building for the Firm, Not Just Himself
At that point, Pranjal recognized a limitation: the scripts worked for him, but they were not scalable. If someone else joined the team, or if he moved on, the automation would be lost. He took the extra step of wrapping the scripts in an HTML front-end, creating a simple interface where anyone could drop in the raw files and get the cleaned output without needing to run code.
The result was a tool the whole firm could use, not just a personal shortcut. What had taken seven days now took roughly fifteen minutes. The time freed up went directly into analysis and insight work, which was the output the client actually paid for.