Add notebooks to fill the basic CLV grid

[ricardoV94]

The idea is to write a notebook with pure PyMC model(s) for each of these CLV scenarios. We can start with the Lifetime part (not-value yet), but ideally we will include value as well by the end. This might be a constant with time-decay penalty in the simplest cases.

Hopefully this will give us a good picture of the building blocks that are necessary for a minimum viable package, and can also serve as the base documentation. Overtime we would replace the custom PyMC code with imports from the CLV sub-package.

• Continuous Non-contractual Adding a distribution class for the continuous non-contractual setting #16
• Continuous Contractual Adding a distribution class for the continuous contractual setting #36
• Discrete Non-contractual (up for grabs)
• Discrete Contractual sBG model notebook for discrete/contractual setting #32

[drbenvincent]

I'll work on adding a notebook for the sBG model for discrete/contractual settings.

[larryshamalama]

I should be able to write something about the continuous contractual setting! I'm feeling inspired 😅

[larryshamalama]

I'm making progress on ^... but I'm knocking my head on a sampling error. Hopefully, I create a PR soon!

[cluhmann]

Is the goal to add more than one model to each quadrant? The CNC quadrant now has @larryshamalama 's BG/NBD distribution (#16), but there are several other models available in that quadrant. I wasn't sure if the idea was to get one per quadrant max or if the goal was broader than that. Also, the distribution added in #16 is named ContNonContract, which seems to imply that there aren't going to be other continuous non-contractual distributions.

[larryshamalama]

Is the goal to add more than one model to each quadrant? The CNC quadrant now has @larryshamalama 's BG/NBD distribution (#16), but there are several other models available in that quadrant. I wasn't sure if the idea was to get one per quadrant max or if the goal was broader than that. Also, the distribution added in #16 is named ContNonContract, which seems to imply that there aren't going to be other continuous non-contractual distributions.

Sorry for not replying to your question initially. Effectively, there are other continuous non-contractual distributions, as I realized later on (e.g. ParetoNBD). We briefly touched upon this point in #128

[cluhmann]

Discrete contractual is covered by #133. Looks like a PR to address #176 would cover the final quadrant.

[ColtAllen]

We have a distribution block for Continuous Contractual, but no modeling options per #279. I've gotten questions on LinkedIn about modeling purchases made through active memberships and/or phone app sessions, so there's demand for it.