#19 Analytics

Closed
opened 2 years ago by jyamihud · 2 comments

In Odysee there in an analytics tab that I use quite rarely. Since it has almost no valuable information about anything at all. Also there is the whole block-chain, with the wallet history and other things that could be used as a data for analytics.

As more people watch a given publication, more LBC is being supported to it. And we can use this data to extract a graph that could be very useful to a lot of LBRY users.

I know that it doens't sound like a terminal feature. But a simple graph in a terminal is implementable. Also using something like pycairo we can generate a more precise image, documenting a given graph. Which could be saved as a file and opened in a default image viewer application.

A discussion of how this could be done, could happen here. I would like to see your ideas.

In Odysee there in an analytics tab that I use quite rarely. Since it has almost no valuable information about anything at all. Also there is the whole block-chain, with the wallet history and other things that could be used as a data for analytics. As more people watch a given publication, more LBC is being supported to it. And we can use this data to extract a graph that could be very useful to a lot of LBRY users. I know that it doens't sound like a terminal feature. But a simple graph in a terminal is implementable. Also using something like pycairo we can generate a more precise image, documenting a given graph. Which could be saved as a file and opened in a default image viewer application. A discussion of how this could be done, could happen here. I would like to see your ideas.

With Commits:

... the analytics is half implemented. You can already tell with some precision what stuff is more popular and when people were watching / reading. But there are problems to still fix:

  • Access to raw data for the analytics. So far you can see on a graph the start and the ends time codes. And you have to infer from the image when a certain spike happened. Perhaps a way of interacting with spikes could be done. Either as a boring list of all entries. Or something a bit more clever. Like having a code for each vertical column of the graph. Which let's you access the list of only the entries in that column.
  • Search in Analytics. For now I need to look through a very slow loading list of publications to access analytics of a certain publication. Which makes it very hard to see analytics of very old publications. Perhaps a way to solve it, is to load analytics from a resolved publication directly. So in url.py we can add sales and analytics options for any resolved publication. That will just draw the graph.
  • Analytics / Sales for the entire Channel / Account. I would like to be able to see a graph for the entire history of all publications. This will require caching all transactions of any given channel, or all transactions, excluding personal transfers and the reward program.
  • Exporting analytics. An export of analytics into some easy to read analytics data files could be done. Like CSV. So a better graph could be generated from a dedicated software.
  • Caching / Saving. Analytics takes time to load. So a way to save the full report into Json and then re-open it right within the graph should exist.
With Commits: - [Analytics! Real analytics. Damn!](https://notabug.org/jyamihud/FastLBRY-terminal/commit/9b5cc8723229a8c251587e6b3c4fc9b54c62d812) - [Added Analytics into Help](https://notabug.org/jyamihud/FastLBRY-terminal/commit/24e726c28b20bdf369cb01c36cad19f608d21a69) - [Softer Graphs. Antialiasing, so to speak.](https://notabug.org/jyamihud/FastLBRY-terminal/commit/fbd00632f51c8ed9fb4ec6b836296a95f1a1cc1c) - [Settings for ASCII graphs and for more characters in tables.](https://notabug.org/jyamihud/FastLBRY-terminal/commit/fcd1b8916f3d3d3cb22a1c42466f8142b1943969) ... the analytics is half implemented. You can already tell with some precision what stuff is more popular and when people were watching / reading. But there are problems to still fix: - [x] **Access to raw data for the analytics.** So far you can see on a graph the start and the ends time codes. And you have to infer from the image when a certain spike happened. Perhaps a way of interacting with spikes could be done. Either as a boring list of all entries. Or something a bit more clever. Like having a code for each vertical column of the graph. Which let's you access the list of only the entries in that column. - [x] **Search in Analytics.** For now I need to look through a very slow loading list of publications to access analytics of a certain publication. Which makes it very hard to see analytics of very old publications. Perhaps a way to solve it, is to load analytics from a resolved publication directly. So in `url.py` we can add `sales` and `analytics` options for any resolved publication. That will just draw the graph. - [x] **Analytics / Sales for the entire Channel / Account**. I would like to be able to see a graph for the entire history of all publications. This will require caching all transactions of any given channel, or all transactions, excluding personal transfers and the reward program. - [x] **Exporting analytics**. An export of analytics into some easy to read analytics data files could be done. Like CSV. So a better graph could be generated from a dedicated software. - [x] **Caching / Saving**. Analytics takes time to load. So a way to save the full report into Json and then re-open it right within the graph should exist.
Sign in to join this conversation.
No Milestone
No assignee
1 Participants
Loading...
Cancel
Save
There is no content yet.