During the PhD · Build your proof

Your technical toolkit

Tools make you capable. Proof that you can use them makes you employable. Here is what to learn, in what order, and how to show it.

3 min read · By Marc Sleiman

Tools make you capable. They do not, on their own, make you employable. The students who struggle are often not the ones who lack skills; they are the ones who list twelve tools on a CV and cannot use any of them under pressure. This page covers what to learn, in what order, and how to prove you can use it.

The foundation

Almost every industry role for an economist expects some version of the following. Treat it as the floor, not the ceiling.

Excel and PowerPoint

Students laugh at this until their first week on the job. Your manager will send you a spreadsheet and expect formulas, pivot tables and a clean chart within the hour. Your findings will travel through the company as slides. Being fluent and fast in both is not glamorous, but it is noticed immediately.

Python or R

Both are widely used in industry, and Stata is not. If your research is in Stata, translate one project into Python or R. You will learn the language and end up with a sample you can show. In Python, the libraries that matter most for economists are pandas for data, statsmodels and linearmodels for econometrics, scikit-learn for machine learning, and matplotlib or seaborn for charts.

SQL

If you learn one new thing this year, make it SQL. Data in companies lives in databases, and you will be expected to pull it yourself. SQL also appears in a large share of technical interviews. See the coding and SQL page for what is tested.

Git

Git tracks changes to code and lets teams work on the same code safely. You need to be able to commit, branch, push and open a pull request. It also puts your work on GitHub, which becomes part of your portfolio.

Power BI or Tableau

Dashboards are how much of a company consumes data. Learn one well enough to build a clean, interactive dashboard from a real dataset. Your university library likely has a data visualization guide and free access to one of them.

How to learn without wasting a year

Let your target jobs decide what you learn. The job postings exercise tells you which tools appear again and again for the roles you want. Learn those properly and ignore the rest until you have a specific reason.

Use a structured course to get the basics, through Coursera, LinkedIn Learning or your university’s own offerings, but do not stop at the certificate. The moment you know enough to be dangerous, start a small project with real data. You learn a tool when you get stuck using it, not when you watch someone else use it.

Practice in short, frequent sessions. Thirty minutes a day of SQL problems for two months does more than a weekend bootcamp.

AI skills: build something, do not just talk about it

Every candidate now says they use AI. Very few can show something they built with it. That gap is your opportunity.

Learn to work with tools like ChatGPT and Claude, with Python, with APIs and with no-code automation tools. Then build a small working system. For example: an agent that collects new Federal Reserve releases, earnings reports or policy announcements as they are published, stores them in a structured database, summarizes each one and flags the items that matter for a particular industry. It does not need to be large. It needs to work, and you need to be able to show the output.

When you describe it, focus on the problem it solves and the result, not the technology. “I built a tool that turns every Fed release into a two-paragraph brief for a bank’s pricing team within minutes of publication” is a sentence an employer remembers.

Show it, do not list it

For every tool you put on your resume, you should be able to point to something you made with it. If someone asks “how have you used Tableau?” the best answer is a link. If you cannot point to anything yet, either build something small or take the tool off your resume until you can.