By using this website, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Customer storiesPricingAboutCareersGet started
href="javascript:history.back(1);" class="w-button">Back Button
← Back
November 29, 2023

How to leverage python + Harmonic to supercharge your sourcing

Aya Spencer

Hello! My name is Aya Spencer and I am a senior analyst at Kapor Capital — an early-stage impact fund based in Oakland California. Prior to being in VC, I worked in various roles within the public and private sectors leveraging analytics to uncover unique trends.

I wanted to share how you can apply insights from analytics to create custom and powerful filters inside Harmonic as a way to maximize your sourcing strategy.

*Disclaimer: I am not an investor in Harmonic at the time of this article. I just really like the platform!

A Bit About Python

If you are ever interested in furthering your data analytics capabilities, I highly recommend learning Python. There are so many things you can do with this language using just a few lines of code.

For example, CB Insights maintains a list of all the unicorn companies worldwide and some key variables along with them. Using Python, I can scrape this information into a dataframe to begin my analysis. After the standard cleaning and wrangling of my data, I can begin my exploratory data analysis (EDA) which ends with a graphical representation of my findings.

Here is an example of a treemap that shows the total aggregated valuations of unicorn companies, broken down by country. As you can see, the United States dominates in aggregate valuations of unicorn startups, followed by China.

In addition to the example shown above, EDA of this particular dataset can help uncover many other insights, including:

  • Which sector produced the most unicorns thus far in each country?
  • Who are the top 5 investors with the highest number of unicorns (per sector)?
  • What are the top 5 sectors with the highest count of unicorns (per city)?

And many others!


Once you’ve uncovered some unique insights from your exploratory data analysis, the fun part is to apply this new knowledge as filters in Harmonic. Harmonic is a database of startups and people (founders, investors, etc.). Their data engine allows for deep insights that you can leverage to source interesting companies. My favorite thing about Harmonic is that you can create custom filters and even have your query results sent to you as a scheduled alert! Let me walk you though a quick example:

I have a personal interest in the Japanese startup scene. Let’s say that I come across a dataset and run some analytics to uncover characteristics of historically successful startups in the Japanese market. My next step would be to turn these insights into filters in Harmonic. I’ll do this by first creating a new search and calling it “Future Japanese Unicorns.”

If my EDA had revealed that most successful Japanese startups are often in the health and wellness space, I can create a filter to search for early stage companies with tags relating to health and wellness.

You can imagine just how powerful these filters can become, especially if you know what signals to look for. You can even “group” your filters together, which becomes really useful if you want to apply multiple conditional statements (ex. Using a combination of “and/or” conditions).

Finally, once I’ve set my desired filters and my query yields promising results, I can automate this search as an alert in Slack.

As a junior VC, I look through tremendous amounts of deal flow each day. Having a scheduled query generated in Harmonic helps me to ensure that I do not miss out on unique opportunities.


I hope this quick tutorial was insightful. Sourcing — especially for early stage companies — is oftentimes an art. Leveraging a quick EDA prior to using Harmonic can help to make the most of your sourcing strategy. As a quick recap, here are the steps to maximize your sourcing strategy:

  1. Use EDA to find interesting trends and signals
  2. Apply these signals as filters in Harmonic
  3. Set your alerts to send scheduled queries to your Slack.

Thank you for reading. Feel free to check out for more of my writings!

Note: this was originally published on Aya’s medium page, give her a follow if you want to read more from her!

Aya Spencer
← Previous Article
There is no previous article
There is no next article
Next Article →