Eod bulk api for daily screener

Hello,

I would like to use the eodhd bulk api for retrieving and storing eod daily stock data for making my own volatility screener (on daily close).

The « bulk » part of the eod api will therefore be very practical, especially for the initial download (need approximately 5 years of daily data for every Nyse stock).

It will then be easy to just download the eod data for each day.

However, i do not see a kind of « diff » command that returns the eod data for the stock that have been subject to an adjustment due to split, etc. It used to be something available on Quandl but it seems I cannot find an equivalent.

As you can guess, the diff is very practical because it enables to avoid to redownload the full set every day.

Is there something like this?

Thanks

Hi! Perhaps a solution would be to get data on all splits and dividends for a particular day using the Bulk API for Splits and Dividends, and then process data only for tickers that had splits and dividends (if I understand your task correctly).

@David_Kellogan The correct way to do this is to download splits and dividends as suggested. The values you get in OHLC are not adjusted and then you need to adjust it yourself. Don’t use adjusted value as that would change on the next split and you can’t be downloading full historical on every split (I mean you could but it’s not practical). You will also want to adjust for dividends optionally and that changes the time series on every dividend.

My strategy is to grab the EOD data for the day, then any tickers with a dividend or split I also grab their full historical EOD data. Combining these together creates a diff similar to what Quandl provided, and I update my stored data with the new diff. I also set up a separate job that grabs 5% of the tickers in my database and refreshes them with full historical data even if they didn’t have a split or a dividend. That way, if there is a data issue, it doesn’t hang around forever. I will get the equivalent of an entirely refreshed database at least monthly.

1 Like