Oct 14, 2021
Supports Http Connections
Build simple analytics apps with columns.ai
Columns Ai is missioned to enable our users to analyze and visualize whatever data they see. Hence, supporting a wide range of data connections is a keep-going effort our builders will do.
We off started supporting Google spreadsheet and CSV data upload, now we have HTTP joined the growing list, public/unauth for now. Even further, thinking columns.ai as a platform, it’s nice to introduce “template” concept, with template, columns.ai transforms a data as an app!
Now, let’s take a quick look at two examples to demonstrate how interesting this could be, the first example is connecting to a REST API returning json payload, the other one is connecting to a normal raw csv file hosted in Github.
Example 1: A simple app to analyze stock data
In this example, we use the REST api provided by ploygon.io, you can experience it now by going to this link: visualize a stock, by entering the stock symbol, it brings up the 2yr data points till query time on columns.ai server for you to analyze and build visualizations from. A story could look like this:
Now, let’s see steps 1 → 2 →3 to create a data as an app:
- Step 1
Sign in columns.ai and navigate to dashboard/data page, click +CREATE and choose “HTTP”, paste an example HTTP link:
Columns previews the payload samples and try to extract rows of data, it also list optional fields for user to define column map as shown in above image.
In this example, it found “results” in the payload contains all the rows, and for each row, we define columns as v →volume, o → open (price), c → close, h → high, l → low and t →time, the unchecked “vw” and “n” are ignored.
- Step 2
Make it dynamic by templating some segments in the URL to a user input or predefined macros — this is the way how you make your data source dynamic depending on final user’s input!
- Step 3
Save the data source by choosing proper column types if necessary and set the TIME column (check its value to match selected time definition). And “import” it.
Now, this new data source should be available in your data collection, as shown like mine:
I shared this data definition, that’s why you can access it from the visual link I shared earlier.
More open data, the world is more connected!
Example 2: import data of raw CSV from Github
The second example, let’s take a look at a Covid data set shared on Github by NYTIMES, this is regularly updated CSV dump at about 100MB, here is the homepage of this data.
Similarly as the first example, just paste the data link after clicked “HTTP” option in the data import page, it detects the data samples and data format as CSV.
Click “submit” and next step clicks “import” after filling the basic data source properties such as name, description.
And we’re ready to query it!
One way I looked at the data
total covid cases ranked by state in 2021
A bit more fancy — I can choose an animation to visualize it, such as:
“total covid cases raced daily by states in 2021”
See this animation story shared publicly on columns:
Conclusion
I hope you enjoyed these 2 examples demonstrated for how HTTP connection is supported.
By supporting HTTP connections, it unblocks many interesting scenarios, when we say Columns Ai is a growing flexible platform that our users can build extensive apps on top of it with unlimited power, this is one good step on track. Cheers!
At last but not least, we would like to hear from you, as builders, we want to build things meaningful to our users, so please get in touch at feedback@columns.ai, cheers!
