in 2022, things were very different.
Children today don't know what it's like.
I used to spend hours:
- Writing Python and SQL code from scratch, line by line
- Remembering which libraries to import and which functions they contain (from sklearn.metrics import r2_score)
- Debug code errors
- Documentation of my code
- Building dashboards to analyze large data sets
Even just in the last year, as AI tools have become more advanced, my work as a data scientist has changed. I'm not a coder and I'm a strategist. Someone who understands the data well in my organization and knows how to best present it and get insights from it.
Claude changes things very quickly
Claude is one of those tools that I believe will change the industry and this profession faster than anyone can imagine. I'm not going to lie, it's scary. At the same time, there are ways that data scientists can take ownership of this tool, master it, and continue to stay ahead of the game.
Here are 3 CRUCIAL skills every data scientist should be working on mastering right now:
1. Claude Dashboards
I used to spend all day building a Tableau dashboard for a client just to test a few questions about a large dataset that might not be looked at again in a few months.
Now, Claude can generate a fully functional, interactive dashboard in minutes, complete with:
- KPI metric cards
- Line charts
- Bar charts
- Drop down buttons
- Tabs
- … and more
Let's show a simple example using the AEP clock data dataset (CC0 license).
Answer from Claude:
I have a dataset of hourly energy consumption time series (AEP_MW) with a time of day column. I created an interactive HTML dashboard that includes:
1. Four KPI cards showing average load, high load, minimum load,
and summer vs winter
2. A line chart showing average load per hour of the day divided by weekday versus weekend
3. Bar chart of monthly average load with peak months highlighted in warm color
4. Bar chart of average load by weekday and weekends in different color. Use a clean, minimal style.
The result looks like this:
A few ideas stand out on the dashboard that wouldn't be available in the raw CSV:
- Weekday usage peaks around 5-6 PM, while weekends peak earlier (around 2 PM) and are generally lower.
- The consumption of July and August is much higher than the spring months, which ensures a strong summer season from the air conditioning load
- Saturday and Sunday loads are always 10% lower than weekdays
These types of dashboards are great for doing EDA and generating one-time reports for stakeholders who just want to know what's going on at the same time. You can also generate a dashboard on a schedule to get a new report every week.
2. Claude Cowork for Prioritizing Jira Tickets and Tasks
Here's what a Monday morning used to look like for me: open Jira, click on 20 open tickets, try to remember the context of each one, figure out what's blocking what, and write a hard priority list for the week.
Claude Cowork is different from Claude Chat in that it actually connects to your desktop and can read/write files. It can connect to Jira (Or other Scrum/Agile platform), and summarize your priorities for the week. Here is an example:
Pull all my open tickets from the current runner. For others, give me: the ticket ID, a one-sentence summary of what needs to happen, the current status, and any blockers. Put yourself first and tell me what I have to deal with first today.
Here are a few other instructions you can use with Cowork:
Writing tickets to Jira
Here are my notes from today's model review meeting: [paste notes – or link to the notes if your Cowork is connected to Google Drive]. Create Jira tickets for each action item in the DS project.
For each one, write a clear, 2-sentence description of what it is
what needs to happen and why, set priorities based on urgency,
and give them a current athlete.
Preparing for the stakeholder meeting
Read the last 3 weeks of comments on tickets marked 'model-deployment' and write me a 5-point status summary to share with the engineering team lead. It ended up being non-technical.
It writes scripts from scratch
Open the preprocessing_pipeline.py file in my project folder and write a README section that explains what the pipeline does, what it expects for input, and what it outputs.
End of sprint reporting
Based on the closed tickets from this runner, write a 3-part running summary for my manager that includes what we posted, what we learned, and what moves on to the next runner.
This is a great time saver and will keep you organized.
3. Debugging with Claude Code
Claude Code is a command-line tool that runs in your terminal with full access to your codebase. It can:
- Read files throughout your project
- Run the commands
- Do the exercises
- Make changes to all multiple files
For data scientists, the most immediate application is to configure pipelines.
Here is a real situation I encountered at work recently with dbt. The names of the models and files have been changed so that I don't share any confidential company information.
I ran dbt run --select fct_energy_forecast and i got this:Database Error in model fct_energy_forecast column "meter_reading_mw" does not exist LINE 14: AVG(meter_reading_mw) AS avg_load_mw,
The problem with dbt models is that the column error in the downstream mart model does not tell you there the column really broke. It may be renamed to the raw source, to the staging model, to the central integration layer, or to the mart itself. To find the cause manually, you would have to open each file in each dependency chain, trace the column name through all the changes, and find out where the old name was not updated. For a project with 24 models and 6 sources, that could be over an hour to read, restart and rebuild the models.
I gave Claude Code instead:
My dbt model fct_energy_forecast fails with 'column meter_reading_mw does not exist'.
Find where this column is defined upstream, trace all dependencies
models and source files, find out what happened, and fix it.
Claude read the entire file from the dependency thread and returned after 40 seconds with the diagnosis.
It then applied the correction to all three lines, re-ran the model, and verified that it was successful.
The conclusion
As tools evolve, so will our roles. Claude is changing the type of work data scientists will end up doing. Instead of spending 8 hours a day fixing various dbt and Python errors, those errors will be resolved in 2 minutes, allowing us more time to dive deeper into our data and ask more important questions. As data scientists in 2026, it is important that we continue to grow our skills and stay up-to-date.
It is also important to note that although Claude has many talents, it's still AI and it can (and does) make mistakes. Data scientists with Claude's skills will still be needed to validate data, improve information and processes, and correct Claude when he's wrong.
