3 ms·

I did this a few years ago with a friend. We settled on using a thing or 2 from Quandl as the best fit (hundreds of $ instead of tens of thousands) after we rea

I did this a few years ago with a friend. We settled on using a thing or 2 from Quandl as the best fit (hundreds of $ instead of tens of thousands) after we realized XBRL was a nightmare. We did this privately for investment purposes; if you're making a user-facing thing, perhaps their terms won't fit your case.

aminozuur 6y ago
I have looked into Quandl, as well as other finance API's. I just had a call with Intrinio, which will offer me a demo soon. I will have another demo from S&P Global, which I fear will be unaffordable, but it's always worth a try. May I ask what specific usecases your internal tool was built for? And were you in the end able to build it to your satisfaction? I welcome any thoughts.
- derivagral 6y ago
  At the time, there was nowhere that we knew of on the net that showed fundamental data and metrics yearly or quarterly in columns and didn't cost many thousands. Google Finance used to sort of do this for 2-3 periods, but the available fields were limited and the UI was not friendly to research. So we made a tool where you could: select various series of data, see their historical series, and also visualize your selections. This was helpful to assist our manual screening after we ran our automated filters (think Morningstar, just as a database query). A real manual screen setup might start with something like: "I want retained earnings, outstanding debt, and fully diluted share counts for the last 5 years for AIZ and their top3 competitors". Plotted out, perhaps you notice that property insurance segments are fairly cyclical (weather/season) which can be a useful thing to tease out for an outsider/amateur. Or maybe you note that companies don't always break their revenue sections out the way you want, so you can't even do this analysis! (Looking at you, Alphabet & Youtube.) We started with Yahoo prices and Yahoo fundamental data, but eventually realized that we still had to verify a lot of their data (against EDGAR). Quandl was a step up in price but we worried a lot less about data integrity with that, and by this time we were seeing some confirmation of results for our strategy. We had a good ARR sustained over a couple years. However, the scalar value of the (risky!) gains was far too small to justify either quitting our tech jobs or sustaining an extra 40+ hours a week. The lessons were great, and seeing shenanigans in real life (looking at you, TSLA) but not in headlines is always fascinating. We did do some concepting of a subscription for the research product, but I'd be surprised if that was realistic given our lack of marketing knowledge (and $). RH is a thing now though, so perhaps we were wrong.

aminozuur 6y ago

I have looked into Quandl, as well as other finance API's. I just had a call with Intrinio, which will offer me a demo soon. I will have another demo from S&P Global, which I fear will be unaffordable, but it's always worth a try. May I ask what specific usecases your internal tool was built for? And were you in the end able to build it to your satisfaction? I welcome any thoughts.

derivagral 6y ago
At the time, there was nowhere that we knew of on the net that showed fundamental data and metrics yearly or quarterly in columns and didn't cost many thousands. Google Finance used to sort of do this for 2-3 periods, but the available fields were limited and the UI was not friendly to research. So we made a tool where you could: select various series of data, see their historical series, and also visualize your selections. This was helpful to assist our manual screening after we ran our automated filters (think Morningstar, just as a database query). A real manual screen setup might start with something like: "I want retained earnings, outstanding debt, and fully diluted share counts for the last 5 years for AIZ and their top3 competitors". Plotted out, perhaps you notice that property insurance segments are fairly cyclical (weather/season) which can be a useful thing to tease out for an outsider/amateur. Or maybe you note that companies don't always break their revenue sections out the way you want, so you can't even do this analysis! (Looking at you, Alphabet & Youtube.) We started with Yahoo prices and Yahoo fundamental data, but eventually realized that we still had to verify a lot of their data (against EDGAR). Quandl was a step up in price but we worried a lot less about data integrity with that, and by this time we were seeing some confirmation of results for our strategy. We had a good ARR sustained over a couple years. However, the scalar value of the (risky!) gains was far too small to justify either quitting our tech jobs or sustaining an extra 40+ hours a week. The lessons were great, and seeing shenanigans in real life (looking at you, TSLA) but not in headlines is always fascinating. We did do some concepting of a subscription for the research product, but I'd be surprised if that was realistic given our lack of marketing knowledge (and $). RH is a thing now though, so perhaps we were wrong.