Does anyone use any APIs (Application Programming Interfaces) for accessing market data so they can do their own data slicing and dicing?

I've been fiddling with it for a few years and have found these to be the most useful - they are free with rate limits and other restrictions, but can still provide enough data for some interesting analysis.

Alpaca Markets: https://docs.alpaca.markets/us/reference...eactions-1
- Has 10 years of dividend announcements - among other interesting things
- Rate limited to 200 calls per minute for free accounts
- I've tried it for getting quotes as well, but found the quotes to be inaccurate. Free quote data comes from the IEX exchange which I think is noted for that.

FMP (Financial Modeling Prep) https://site.financialmodelingprep.com/d...s#calendar
- Limited to 1 month history for free accounts.
- Rate limited to 250 calls per day for free accounts

Unofficial Yahoo Quotes API https://github.com/dshe/YahooQuotesApi
- Doesn't require any sort of account or API keys.
- I've used it for getting smallish (less than 20) batches of quote data without much problem for years.
- Sometimes (rarely) it stops working for a week or so - when Yahoo makes changes - and someone has to fix it.

I briefly looked into a few others and found the free tier for them were limited enough to not be useful.

AlphaVantage - highly rate limited
FirstrateData - Expensive
Polygon - Pay to get anything useful
Tradier - have to have a brokerage account with them
Intrinio - Pay
Xignite - Expensive
Twelve Data - Pre-purchased credits
Schwab - They didn't have an api when I looked originally, I hear they do now though I havn't looked at it again.
Quandi - mostly non-quote types of data. I didn't want to signup for an account
Marketstack - 10 bucks a month.

basicdividends database

I grabbed 10 years of dividend announcement data from the Alpaca service and stuck it in a database - if anyone is interested in trying some of their own SQL Kung-Fu on it. (Note, AI these days is pretty great for generating SQL scripts)

https://drive.google.com/file/d/1D2X_s3e...sp=sharing
* SQLite database (zipped up, its about 4.8MB)
* About 10 years of dividend history
* 16483 unique ticker symbols
* 345308 dividend events

Each dividend event has the following columns:
Symbol, ExDate, PayableDate, DividendRatePershare, IsForeign, IsSpecial

A separate table within the database maps symbol names to payment frequency (monthly, quarterly, etc) where we could figure it out based on the payment history, or directly from another datasource.

Databases are too hard - what can you do with it?

Well, here is one example - I created this report using the dividend database I just posted about above.

Companies or funds that recently raised their regular dividend, while filtering out names with spotty or highly variable dividend histories.

Google Sheet: https://docs.google.com/spreadsheets/d/1...sp=sharing

Each result is the most recent dividend increase for one symbol. The list only includes symbols that:
* Pay monthly or quarterly dividends
* Have at least 5 years of dividend history
* Have not skipped regular dividends during that 5-year period, based on their expected payment schedule
* Have not had any dividend decreases during that same 5-year period
* Recently paid a regular dividend that was higher than the previous regular dividend
* Exclude special dividends, one-time payments, and zero/blank dividend records

The columns show:
* Symbol: ticker
* Frequency: monthly or quarterly payer
* ExDate: ex-dividend date of the increase
* PreviousExDate: prior dividend’s ex-date
* PreviousRate: prior dividend amount
* NewRate: new dividend amount
* IncreaseAmount: dollar increase per share
* IncreasePercent: percentage increase from the prior dividend
* PaymentCountInFiveYears: number of regular dividends found in the 5-year review period
* MaxGapDays: longest gap between regular dividends during that period

The results are sorted by the most recent dividend increase first. If several symbols increased dividends on the same date, the larger percentage increase is shown first.
Note: I arbitrarily capped the results at 1000 - which cuts of the results data in Feb 2022

Important caveat: this is a data-based screen, not a judgment that the dividend is safe or that the company formally announced a dividend policy increase. It simply identifies symbols whose recorded regular dividend payments show a recent increase and a steady 5-year payment pattern.

I can share the SQL code with anyone who's interested - its ugly. I used an AI tool (Codex) to create it

Dividend Growth Forum - All Forums

Message History

Delivery option(s)