Show HN: Chatty – analyze texts to and from your friends | Hacker News Reader

Show HN: Chatty – analyze texts to and from your friends

github.com

11 pointslpage10 comments

README is pretty vague and says it "analyzes" text messages. In what way? There's no examples of the analysis it outputs.

Agreed, it does need a better README. Checking the code, it seems the metrics generated are defined here:

https://github.com/klittlepage/chatty/blob/master/chatty/ana...

So it's mostly character analysis (word and message length, emojis). The interesting two metrics, in my mind, and the percentage of positive and negative sentiments in the messages.

Sorry about that - I'm relying on the notebook itself as the example. If you open it up, you'll see a full analysis.

Nice. I was just downloading manually chats from LINE app to try something like this a few days ago.

I know that there used to be an SQLite database accessible, but it seems to be gone now and I've searched everywhere (incl. hidden folders in iCloud) but it's nowhere to be found. Have you done any research on it?

Also LINE doesn't seem to sync the (historical only, maybe) chats between macOS application and the iOS application - making the history not exactly true. When I can even get the messages that is.

The NLP nerd in me loves this! I'm curious, though. Do you have an agenda? Where do you see this going?

Nope, just an evening hack :)

As for what's next, I'm adding time as a dimension for analysis, entity recognition, and an .mbox source for e-mail. That aside, feature requests and PRs welcome!

I don't have conda installed on my mac so I tried for a bit to get this working in a docker file based on continuumio/anaconda3. But `conda env create` keeps returning an error: ResolvePackageNotFound for appnope 0.1.0 py36_3 (the first package in environment.yml, so probably symptom of a larger problem).

Wasn't able to get it working, but it looks cool.

I'm not sure about that error specifically, but I just simplified the environment.yml file s.t. only direct project dependencies are included. Hope that helps!

Including only direct dependencies without build numbers should take care of the problem.

Hi, author here. Happy to answer any questions. Check out the notebook itself for a sample analysis. If you're on mobile, make sure to request it as a desktop site or GitHub will render it raw.

Reply on news.ycombinator.com