## The 3 most interesting questions this dataset can answer ### 1. Is the Indian startup ecosystem consolidating into fewer, bigger deals? The number of reported deals falls every year while the typical disclosed ticket grows (step-9 query, by year): | Year | Deals | Median disclosed ticket (USD) | |---|---|---| | 2015 | 935 | 1.5M | | 2016 | 993 | 1.0M | | 2017 | 687 | 2.25M | | 2018 | 310 | 4.0M | | 2019 | 111 | 12.0M | - Deal count falls from about 990 in 2016 to 111 in 2019. - The median ticket rises roughly twelvefold over the same period. - Where totals are computable, 2019 raised about $9.7B across only 104 disclosed deals. 2017 raised about $10.4B across 456 disclosed deals. - This bears directly on the question of whether funding is accelerating or slowing, and on how much a founder should expect to raise. - 2020 has only 7 deals, so it is too thin to read. - **Caveat:** the fall in deal count may partly reflect reporting coverage rather than real activity. This is a "reported deals" dataset, not a census. ### 2. Which cities dominate, and how do ticket sizes differ by city? Bangalore leads with 700 deals and about $11.4B disclosed, followed by Mumbai (567) and New Delhi (421), with Gurgaon at 287 (step-10 query). - **Bigger tickets, fewer deals:** median tickets differ noticeably by city. Bengaluru's median is $5M, Gurugram's is $4M and Chennai's is $2.5M. Most other hubs sit around $1–2M. - **Alias problem:** the city column holds aliases for the same place, which splits the true totals: - Bangalore and Bengaluru - Gurgaon and Gurugram - New Delhi and Delhi - **Missing city:** 171 deals have no city at all. - Merging the aliases would make Bangalore's lead even larger. - This informs where to launch, open an office or source deals. ### 3. Which sectors and round types attract capital, and has the mix shifted over time? - The dataset has 822 distinct industry verticals and 57 distinct investment-type values, plus investor names on all 3,044 deals. That supports questions such as "what is the typical round and cheque size for fintech or e-commerce?" and "who are the most active investors by sector or city?" - **Taxonomy change:** the vertical taxonomy is not consistent over time. 2015 has 724 distinct verticals against only 19–53 in later years, which suggests free-text labels in 2015 and a controlled list afterwards. - Any vertical trend therefore needs the labels normalised first, or should be restricted to 2016 onward. ## Data-quality caveats - **Disclosed amounts:** only 2,066 of 3,044 deals have a numeric amount. All sizing conclusions rest on disclosed deals only. - **2016 total:** the 2016 total returned `nan`, so it cannot be summed without cleaning the non-numeric amount strings. - **Dates:** a few dates are malformed. Seven deals have unparseable dates and one parses as year 15, so the date column needs light cleaning before time-series work. ```json {"charts":[{"kind":"combo","x":"yr","series":["deals","median_usd"],"line_series":["median_usd"],"title":"Deals fall while median ticket rises"}]} ```