A number can be calculated correctly and still answer a question the evidence never asked. I want the reader to know what was observed, what was grouped together and which conclusion remains unavailable.
Joseph Nelkashi
Data organisation
29 colleagues
Four heads coordinating observation, research and publication data
Annual function expense
₹5.8 crore
Includes ₹1.6 crore methodology development
Research and data career
33 years
Source records through commercial interpretation and publication accountability
Leadership, seen by others.
Joseph makes the narrower conclusion understandable to a customer, including when it changes a publication he previously approved.
The average rose while selected age-band averages stood still
Two selected groups of twenty transactions produced different averages in a 2023 review. The first contained fifteen older-band vehicles averaging ₹8 lakh and five newer-band vehicles averaging ₹16 lakh. Its overall mean was ₹10 lakh. The second contained ten in each band, with the same respective band averages, and an overall mean of ₹12 lakh. The twenty per cent rise in the raw mean reflected the changed mix in this example; it did not establish a complete constant-quality price index.
I am Joseph Nelkashi, chief data officer of Kershavi Vehicle Intelligence in Kochi. Our 163 colleagues provide commissioned research and subscriptions. Four heads coordinate my twenty-nine-person function, while Farhan leads methods and Anita retains enterprise authority. I want the account to preserve the commercial question without allowing a convenient aggregate to settle a different one. Vehicle age is only one characteristic; the example cannot establish what happened to every comparable vehicle or to the wider market.
That discipline includes an error I approved in 2018. Thirty-six removed listings had been described as completed sales. Only twelve transactions were confirmed; the other twenty-four removals had unknown outcomes. We withdrew the sales claim without treating those unknowns as unsold vehicles. In 2025 we kept bundled deals separate from a single-vehicle benchmark. One twelve-vehicle bundle sold for ₹1.44 crore: dividing the total produced ₹12 lakh per vehicle, but that arithmetic allocation was not evidence of each vehicle’s achieved price. My continuing remit is to make the distinction survive publication, customer explanation and board discussion. These selected fictional accounts demonstrate judgement rather than market guidance. I have no statutory seat and would need further governance preparation before considering an outside appointment, with twenty-three proposed annual days.
Two groups, a different mixture
Twenty transactions in each selected group produced means of ₹10 lakh and ₹12 lakh while the older and newer age-band means remained ₹8 lakh and ₹16 lakh.
The removed listing did not establish a sale
Twelve confirmed transactions and twenty-four unknown outcomes replaced a claim that all thirty-six removed listings represented completed sales.
A bundle total remained a bundle observation
A ₹1.44 crore deal covering twelve vehicles was retained as a combined transaction rather than twelve observed individual prices.
Data authority with a separate methods challenge
Twenty-nine colleagues work through four heads within ₹5.8 crore annual expense, including ₹1.6 crore methodology development.
Current focus
Vehicle-market accounts with visible observation and comparison limits
Conversations I welcome
Data leadership and a prepared first contribution to research oversight
The leadership record
Trusted with consequential decisions.
Formation
Bachelor of Science in StatisticsNerpilvi Statistical College · 1993
