net prem ticks dashboard v2

We want use the net prem ticks and OHLC endpoints to create a thumbnail dashboard for 6 tickers, say a "sector" dashboard like semiconductors. So the first thing we will do, after setting up grabbing our API token and setting up headers, will be a call to the net prem ticks endpoint:

https://api.unusualwhales.com/docs#/operations/PublicApi.TickerController.net_prem_ticks

We can see in the documentation that each record represents the data for a single minute's ticks, and the documentation suggests that in order to build a daily chart we will need to add the previous data to the current record, so we will tackle that after getting the data into a DataFrame for easy manipulation.

Success! If we were to execute net_prem_ticks_rsp.json(), we would get data like this:

{'data': [{'date': '2025-01-23',
   'net_call_premium': '-167229.0000',
   'net_call_volume': -270,
   'net_put_premium': '9946.0000',
   'net_put_volume': -129,
   'tape_time': '2025-01-23T14:30:00Z'},
  {'date': '2025-01-23',
   'net_call_premium': '-18254.0000',
   'net_call_volume': 35,
   'net_put_premium': '21329.0000',
   'net_put_volume': 101,
   'tape_time': '2025-01-23T14:31:00Z'},
   ...
   {'date': '2025-01-23',
   'net_call_premium': '-844559.0000',
   'net_call_volume': -2526,
   'net_put_premium': '-4246.0000',
   'net_put_volume': -68,
   'tape_time': '2025-01-23T20:59:00Z'}]}

Which is perfect for a polars DataFrame. I like using a DataFrame for this kind of information because we can transform and add to the existing data quickly almost as if it were in a spreadsheet. Let's use our DataFrame to:

• Convert strings to numerics or dates
• Add an eastern time zone column and a plot-friendly HH:MM:SS column too
• Calculate the cumulative net premium figure for calls and puts (since each observation only represents that single minute in time)

OK let's go:

Nice! Now the interactive plotting library that I prefer in these notebook settings is lets-plot, and lets-plot prefers "narrow" DataFrames, so we are going to use a method called melt() to create a plot-ready DataFrame:

With that we are ready to plot (with some extra styling)!

Would be cool to see the price of the underlying stock on a similar chart, say right underneath this one, so let's make that happen. We can use the OHLC endpoint:

https://api.unusualwhales.com/docs#/operations/PublicApi.TickerController.ohlc

To get 1-min data and in this case just work with closing prices. We're going to use a lot of the same tactics that we did in the steps above, so the explanations will be a bit shorter over the next few cells:

Success, now when we execute ohlc_rsp.json() we get results like this:

>>> ohlc_rsp.json()
{'data': [{'close': '265.6101',
   'end_time': '2025-01-24T00:56:00Z',
   'high': '265.6101',
   'low': '265.6101',
   'market_time': 'po',
   'open': '265.6101',
   'start_time': '2025-01-24T00:55:00Z',
   'total_volume': 4704743,
   'volume': 144},
  {'close': '265.63',
   'end_time': '2025-01-24T00:54:00Z',
   'high': '265.63',
   'low': '265.63',
   'market_time': 'po',
   'open': '265.63',
   'start_time': '2025-01-24T00:53:00Z',
   'total_volume': 4704599,
   'volume': 2572},
  {'close': '265.94',
   'end_time': '2025-01-24T00:49:00Z',
   'high': '265.94',
   'low': '265.94',
   'market_time': 'po',
   'open': '265.94',
   'start_time': '2025-01-24T00:48:00Z',
...
   'market_time': 'po',
   'open': '266.48',
   'start_time': '2025-01-23T00:00:00Z',
   'total_volume': 6595506,
   'volume': 6595506}]}

So let's apply the same transform and enhance strategy from above:

Now that we have our OHLC data collected and ready, we need create a plot-friendly melted DataFrame:

And now we are ready to create our price plot and bring it together with the net prem ticks plot! Let's rest them right on top of one and other to make it very simple to price and net premium flows on the same time-axis:

Extremely interesting to see SMH price rallying all day while Calls and Puts were net sold, right? My suspicion is that market participants are holding a subset of SMH names and using the ETF as a hedging instrument.

To check this hypothesis, let's create a dashboard of the largest semiconductor names using these exact same tactics, except looping through all the tickers (instead of just working with SMH like we have been doing up to now).

Since we have already done this work once the commentary will be a bit more sparse, but I will add notes as I go along just to make sure everything is clear.

OK great, now we have all of our raw DataFrames, let's clean them up then create "melted" versions for the eventual plots:

And let's repeat this for our OHLC data as well, collecting the raw data for each ticker, cleaning each DataFrame, then melting it for plotting purposes:

We are now ready to pair off our net prem tick and price DataFrames to create our dashboard, so let's take care of that then create our thumbnail-sized dashboard: