>This started as a personal project because I wanted a clean, searchable dataset of startups across regions without jumping between multiple sources or dealing with noise I didn't want :).
I love this idea!
Something like that really needs to be done, and you've stepped up to the plate to begin that journey of putting all of that together!
A list of ALL startups, in one place, would really be great!
One question/caveat though -- how do you determine / how would you determine if a startup is no longer in startup mode?
That is, if the startup has become a big business, if the startup has been acquired, if the startup has failed, etc., etc.?
I guess (if the correct data wasn't present or unavailable or hard to parse, etc.) you could simply take startups off the last after a fixed time period, like maybe 12 months, 24 months, <?> months, ?.
Or, maybe add a retrieval date and source...
Two extra fields for your database... the date when it was spidered/sourced/parsed/found/uploaded/etc, and the source URL (or URL's...).
Then you could keep all of the data for all time... just let your users sort/filter on that retrieval date, for "freshness" of data, relative to their needs...
Hey Peter, Thanks! I really appreciate the thoughtful feedback and your time.
> how do you determine / how would you determine if a startup is no longer in startup mode?
It is a challenge as startups transition all the time in different ways - funding rounds, IPOs or the dreaded deadpool - and I'm trying to figure out the best way to represent it. At this time, I'm using a combination of manually vetting, to soliciting public feedback through "edit this profile" button and showcasing the latest state. Rather than deleting entities that are no longer startups, I tag them with statuses like Public, Acquired, Shut Down, etc., and surface that on the profile page. Here is an example, https://startups.in/united-states/airbnb (you can find a badge under the logo and if you scroll down you can see a card that show the exit details).
As can be seen, Airbnb is marked as a "Public Company" with IPO metadata (ticker, exit date, exit value), and still remains in the database as part of the ecosystem rather than disappearing. The current idea is to treat this more like a longitudinal startup graph.
Long-term, I'd like this to behave more like a "historical record" of startups over time (dare say wikpedia for startups but presented differently?), not just a snapshot of "current startups". That way acquisitions, failures, and IPOs become first-class signals instead of reasons to delete data. Thanks again.
I'm mostly using this as a learning project around data aggregation and enrichment at scale. The biggest open problems so far have been data quality, entity resolution, and keeping things up to date.
>This started as a personal project because I wanted a clean, searchable dataset of startups across regions without jumping between multiple sources or dealing with noise I didn't want :).
I love this idea!
Something like that really needs to be done, and you've stepped up to the plate to begin that journey of putting all of that together!
A list of ALL startups, in one place, would really be great!
One question/caveat though -- how do you determine / how would you determine if a startup is no longer in startup mode?
That is, if the startup has become a big business, if the startup has been acquired, if the startup has failed, etc., etc.?
I guess (if the correct data wasn't present or unavailable or hard to parse, etc.) you could simply take startups off the last after a fixed time period, like maybe 12 months, 24 months, <?> months, ?.
Or, maybe add a retrieval date and source...
Two extra fields for your database... the date when it was spidered/sourced/parsed/found/uploaded/etc, and the source URL (or URL's...).
Then you could keep all of the data for all time... just let your users sort/filter on that retrieval date, for "freshness" of data, relative to their needs...
Anyway, looks great so far!
Great work!
Hey Peter, Thanks! I really appreciate the thoughtful feedback and your time.
> how do you determine / how would you determine if a startup is no longer in startup mode?
It is a challenge as startups transition all the time in different ways - funding rounds, IPOs or the dreaded deadpool - and I'm trying to figure out the best way to represent it. At this time, I'm using a combination of manually vetting, to soliciting public feedback through "edit this profile" button and showcasing the latest state. Rather than deleting entities that are no longer startups, I tag them with statuses like Public, Acquired, Shut Down, etc., and surface that on the profile page. Here is an example, https://startups.in/united-states/airbnb (you can find a badge under the logo and if you scroll down you can see a card that show the exit details).
As can be seen, Airbnb is marked as a "Public Company" with IPO metadata (ticker, exit date, exit value), and still remains in the database as part of the ecosystem rather than disappearing. The current idea is to treat this more like a longitudinal startup graph.
Long-term, I'd like this to behave more like a "historical record" of startups over time (dare say wikpedia for startups but presented differently?), not just a snapshot of "current startups". That way acquisitions, failures, and IPOs become first-class signals instead of reasons to delete data. Thanks again.
I'm mostly using this as a learning project around data aggregation and enrichment at scale. The biggest open problems so far have been data quality, entity resolution, and keeping things up to date.