**Yes. The most natural segmentation uses four dimensions: funding round type, city, industry vertical and year. Each needs its labels cleaned first.** **1. Funding round type (`InvestmentnType`)** - Two groups dominate. "Seed Funding" has 1,385 deals and "Private Equity" has 1,357 (step 3). - A long tail follows: Seed/Angel variants, Debt Funding (25), Series A–F and a few one-offs such as Crowd Funding and Maiden Round. - The raw labels are messy. Seed/Angel appears at least five ways: "Seed/ Angel Funding" (60), "Seed / Angel Funding" (47), "Seed/Angel Funding" (23), "Angel / Seed Funding" (8) and "Seed / Angle Funding" (3). - Other variants include "Seed\nFunding", "Seed Round", "Private Equity Round" and "Pre-Series A" / "pre-Series A". - Collapsing these into about five stages gives a clean grouping: **Seed/Angel, Private Equity, Series A–H (venture rounds), Debt, and Other**. **2. City (`City Location`)** - Deals are concentrated in a few metros: Bangalore (700), Mumbai (567), New Delhi (421) and Gurgaon (287). Pune, Hyderabad, Chennai and Noida each have about 90–105 (step 1). - The column has 944 distinct raw values, but most of that is noise: - Aliases: Bangalore and Bengaluru (141), Gurgaon and Gurugram (50), New Delhi and Delhi. - Non-breaking-space variants such as "\xa0Gurgaon". - Typos such as "Ahemadabad". - Multi-city entries such as "Pune / US" and "Bangalore / SFO". - Missing values stored as "nan" (171), blank (6) and "N/A" (3). - After merging aliases, a sensible grouping is **the top metros, other Indian cities, and international or multi-city**. **3. Industry vertical (`Industry Vertical`)** - This is dominated by Consumer Internet (941) and Technology (478), followed by eCommerce, Healthcare, Finance, Logistics, Education and Food & Beverage (step 2). - eCommerce is spelled several ways ("eCommerce" 186, "ECommerce" 61, "E-Commerce" 29), and 171 values are "nan". - The column has about 903 distinct values, so use the top 10–15 verticals and group the rest as "Other". **4. Time (`Date dd/mm/yyyy`)** - Deals by year: 2015 (935), 2016 (993), 2017 (687), 2018 (310), 2019 (111) and 2020 (7) (step 2). - Activity peaked in 2016 and then declined. - The date column is stored as text, and some values contain non-breaking-space artifacts, so it needs parsing before use. - Year, or year-quarter, is a natural cohort axis, and it can be crossed with round type or city. **Practical recommendation:** Segment first by round type and year, then by city tier and top vertical. Standardise the labels before grouping, because otherwise duplicate spellings split what is really one segment. I did not examine deal size by segment. That would be the next step, and it would require parsing `Amount in USD`, which is stored as text.