birds eye review

[Zethson]

Comments on https://lamin.ai/docs/birds-eye

Generally, I already really like it.

• [x] As you know, I prefer slightly longer introductions, but we generally want to keep them short. However, here I think that we might have to introduce the term "data lineage" in 0.5 sentences. The "why should I care about data lineage" section is so great that I'd actually not make it a drop down here (though I love these in the other sections). I'd also introduce the dataset in 0.5 sentences. I'm confident that with 2-3 sentences we can cover all of this
• [x] "It’s a mix of a guide & a demo usecase." -> I'd remove this.
• [x] This is pretty long and the "assert you're testuser1" and "login with testuser2" is just noise. Let's try to somehow get rid of this. Ooooor we are very explicit that we want to simulate multiple users in this notebook to demonstrate collaborative usage
• [x] The new lines seems off here to me

  
  # app upload
  ln.setup.login("testuser1")
  transform = ln.Transform(name="Upload GWS CRISPRa result", type="app")
  ln.track(transform)

upload and analyze the GWS data

filepath = ln.dev.datasets.schmidt22_crispra_gws_IFNG(ln.settings.storage) file = ln.File(filepath, description="Raw data of schmidt22 crispra GWS") file.save() ln.setup.login("testuser2") transform = ln.Transform(name="GWS CRIPSRa analysis", type="notebook") ln.track(transform)

file_wgs = ln.File.filter(key="schmidt22-crispra-gws-IFNG.csv").one() df = file_wgs.load().set_index("id") hits_df = df[df["pos|fdr"] < 0.01].copy() file_hits = ln.File(hits_df, description="hits from schmidt22 crispra GWS") file_hits.save()

I'd go for

app upload

ln.setup.login("testuser1") transform = ln.Transform(name="Upload GWS CRISPRa result", type="app") ln.track(transform) filepath = ln.dev.datasets.schmidt22_crispra_gws_IFNG(ln.settings.storage) file = ln.File(filepath, description="Raw data of schmidt22 crispra GWS") file.save()

upload and analyze the GWS data

ln.setup.login("testuser2") transform = ln.Transform(name="GWS CRIPSRa analysis", type="notebook") ln.track(transform) file_wgs = ln.File.filter(key="schmidt22-crispra-gws-IFNG.csv").one() df = file_wgs.load().set_index("id") hits_df = df[df["pos|fdr"] < 0.01].copy() file_hits = ln.File(hits_df, description="hits from schmidt22 crispra GWS") file_hits.save()

- [x] "Let’s see how the data lineage of this looks:" -> Let's see what the data lineage of this looks like
- [x] Add a new line above scanpy & move it 1 line down because it concerns the next section in the hidden code cell

screen_hits = file_hits.load() import scanpy as sc

- [ ] Can we show the plot? Would just be fun to look at after there have been so many steps in the data pipeline
![image](https://github.com/laminlabs/lamin-usecases/assets/21954664/ca12be47-d59a-4ae2-bca4-7f259ddc633c)

[falexwolf]

One comment on this:

I'm generally agreed! 0.5 sentences on what data lineage is would be awesome. If everything stays at 2 - 3 sentences, we can also eliminate the dropdown!

One remark: Let's please try to use the same sentences/words/definition that we use in the tutorial.

The idea is that most people will read the readme/landing & tutorial and then come here when they want to know more about data lineage. These people both will have a rough idea of data lineage already and they'll also have seen LaminDB's take on it in the tutorial and on the readme/landing.

A few people might come here without first reading the tutorial. For them the 0.5 sentences on what data lineage is is gold.

But in any case, language should be ultra-consistent between different sections of the docs.

Soonish, I'd also suggest to add both data lineage and provenance to our glossary.

[falexwolf]

Really 100% agreed on everything else! @sunnyosun & I are now trying to come up with a video that actually walks through this using LaminDB & LaminApp.

[falexwolf]

If you prettify anything, please start with this one!