corradio / polynomial Goto Github PK

• Use LinkedIn API to download avatar?
• See https://github.com/bennylope/django-organizations (has invitations)
• https://django-invitations.readthedocs.io/en/latest/
• User with verified domain suffix should automatically join?

When pressing save, ask if backfill (will overwrite previous points)
"This metric is able to fetch past data, would you like to fetch up to X days"?

In addition to this, another thing I was thinking about when setting up the plausible integration is that it would be cool if one could see the graph of daily visitors and also see where they come from, e.g. have it all in one chart or in separate charts. Right now (as far as I understand) I can filter for specific sources but can't create a graph showing the number of people coming from each specific source.

it would be cool to potentially create different views, e.g. have one view focused on marketing-specific insights where a selected number of KPIs is shown and then a second one focused on sales. This would help teams with many KPIs to create a nicer overview

It would be nice to create different dashboards, once you add 20+ metrics the scrolling becomes a bit long. Having the option to create a marketing-focused one or sales-focused etc would be nice to make the overview clearer

Mailgun is not activated

• open source only integrations, and potentially a bit of django to be able to test
• Remove all storage
• Consider importing as django app with poetry github instead of submodule

• Question: should a metric belong to a user or to an org?
  • what happens when a user is deleted? Should the metric still exist?
  • should all metrics of a user in an org automatically "belong" to the org?

• Subscribers (daily)
• Revenue (daily)
• MRR (daily)

https://stripe.com/docs/stripe-apps

Internal Server Error: /metrics/new/UodAM23XrDpNQDq7jsgHzDKRejSa-b1pSxfiMGRM4c0/

IntegrityError at /metrics/new/UodAM23XrDpNQDq7jsgHzDKRejSa-b1pSxfiMGRM4c0/
duplicate key value violates unique constraint "unique_metric"
DETAIL: Key (user_id, name)=(2, Leads number) already exists.

Notes:

• Consider using https://whitenoise.evans.io/en/stable/ for static files

It would be terrific if we could e.g. pre-populate the list of viewIds of Google Analytics right after a user is authenticated.
This would mean be able to generate the schema based on a method which would do a GA request, or do update the schema in the __enter__ section of the integration.

An outstanding question is how to update the rendered schema if it changes dynamically.
Is it possible to update it client-side?

• what keywords would I search for?

What strategy should we use? What should be sent to admins vs what should be sent to users?

• Introduce a UserException class which represents errors that can be sent to the user? What about 401?
• handle_integration_error?

List of exceptions that can be sent to the user:

• 401 requests
• OAuth errors, ask for re-auth

We don't necessarily need it right now, but if we continue the project we will in order to implement:

-> Email confirmation
-> OAUTH

Since it messes with the User model, it's probably wise to include it from the beginning to avoid trouble later.

Imported from spreadsheet, error on /:

Uncaught SyntaxError: Unexpected token 'N', ..."{"value": NaN, "date"... is not valid JSON
    at JSON.parse (<anonymous>)
    at (index):688:21

• api_calls_daily
• api_daily_active_users
• api_latest_delay_seconds
• api_latest_measured_zones_available
• api_latest_zones_available
• api_tokens_valid
• app_visitors_daily
• app_visitors_dk_daily
• app_visitors_us_daily
• blog_visitors_daily
• co2signal_daily_active_users
• forecast_carbon_intensity_mae
• forecast_carbon_intensity_r2
• forecasts_carbon_intensity_mae
• forecasts_carbon_intensity_max_mae
• pipedrive_open_deals
• twitter_mentions_daily
• website_visitors_daily
• website_visitors_dk_daily

• Only show last 30 days
• Make sure last date is today (and not now())

• api_calls_daily
• api_daily_active_users
• api_latest_delay_seconds
• api_latest_measured_zones_available
• api_latest_zones_available
• api_tokens_valid
• app_visitors_daily
• app_visitors_dk_daily
• app_visitors_us_daily
• blog_visitors_daily
• co2signal_daily_active_users
• forecast_carbon_intensity_mae
• forecast_carbon_intensity_r2
• forecasts_carbon_intensity_mae
• forecasts_carbon_intensity_max_mae
• pipedrive_open_deals
• twitter_mentions_daily
• website_visitors_daily
• website_visitors_dk_daily

See https://developers.facebook.com/docs/instagram-api/guides/insights/

Right now it uses an API Key. It would be simpler to use OAuth2 instead.

Number of deals in stage X

To reproduce:

• http://localhost:8000/metrics/5/

TokenExpired happened, and token renewal didn't work
During handling of the above exception ((token_expired) ), another exception occurred:
...

Final error: InvalidGrantError

There always needs to be daily data for the tool to catch the dates. If you want to include monthly data, you need to split out the months into separate days then it is possible to add the data. It would be nice to have an option with monthly data.

Collect views from YouTube channel

• It would be great to have the option to share the dashboard with a team, here we could either have it as just "view" or "edit", edit would allow everyone to add new metrics which can be great but can also become messy faster

• Probably Celery is the best tool for this
• I guess hourly, daily & weekly frequencies are good enough to start (monthly is problematic because month isn't a fixed time interval)

Sentry etc..

e.g. force people to authenticate before creating/editing an integration

Sorting works alphabetically right now which is okay as you can trick it quite easily but it would be nice to have the option to drag and drop (or through a different way) sort the metrics as you wish

Adblocker seems to block calls that include googleanalytics in their URI.
User is currently not notified of failure.

Ability to use OAuth as authentication mechanism in integrations.
This is required for Google products, and will enable marketing analytics such as LinkedIn.