I really like the idea of tracking myself (e.g. books read, weight, steps taken etc.), but as a hacker I'd never let anyone else hold this data. I mean, Facebook is one thing, but highly personal details about my personality is another thing.
That said - it would be cool to see something like this open sources. Currently I'm using excel to track most of the stuff I'm interested in, but I know that there must be a better alternative.
I totally understand. I'm the kind of person who keeps a diary inside an encrypted true-crypt folder, so security and privacy have been my primary concern from day one. The main reason I'm deploying this as a web-service rather than a downloadable app is so that you can track things from anywhere (without having to be technical enough to manage your own servers). It's much harder to do that with excel or true-crypt, even if you use it with things like google-docs or dropbox.
Still, and I don't mean to bash your work, I think this is a disservice to the internet because it perpetuates cloud tools and keeps the state-of-the-art even for something as personal as journaling in the cloud. That's a very, very sorry situation. Paraphrasing the old adage, if you haven't something to contribute that's in the best interests of people, better not to contribute at all. I keep my private stuff local as well, and you yourself admit to doing the same. Why not focus on making it easier for people to do the same, instead of riding the popularity contest of the quick-start cloud?
Because I believe that privacy and the cloud can co-exist in a mutually beneficial fashion. Just because the app is online doesn't mean that it has to be "social" (there is intentionally no social-sharing options for the log or journal entries). But by being online, the application itself can improve over time by analyzing aggregate data to detect anomalies (like Mint.com), give predictions, refine the natural language parsing etc.
If this was just a journal, I would agree with you and make it self-contained. But it's not. In the next 5-10 years we're going to be seeing a massive number of wearable sensors embedded in our clothes, appliances and tools. A managed, online API will be better able to be always available, consuming this tsunami of data and turning it into useful insights for you.
How do you track things you're interested in Excel? For instance, if you're into python, would you make a python worksheet and fill it with goals and things to read?
I think what he means is that he tracks metrics that he wants to track (such as steps walked, miles run, water consumed, calories consumed, blood pressure, lines of code, dollars spent, etc.) in Excel.
You could track things related to your programming such as lines of code, optimizations made, errors encountered, bugs fixed, refactoring done.
By the way, the site's running entirely on a single ec2 micro instance (nginx, redis, posgres and go). I'm curious to see if my optimizations pay off and if it can continue to handle the HackerNews traffic.
That looks pretty good! You have an interesting model of abstraction with the records, cats and magnitudes. For me, each data point is an Entry. Entries have a primary Metric/Action and many details, which may have a numerical amount associated with them. Free-form representations like yours and mine make logging data easy, but analyzing it can get pretty tricky.
This is executed very well. I love the natural language processing for diary entries.
However I think you need to change your pitch if you want more incoming traffic. At the moment this has an aura of creepiness and vainness to me and no chance of SEO (who googles for "track every moment of my glorious life in detail"?). You should configure this for whitelabelling and create a version for different markets.
This is just one idea because I knoe my wife would love it:
Dieters to track calories eaten each day and exercise done. Would be awesome to see energy counts of food eaten vs energy consumed by exercise done - maybe like a weekly comparison of energy in and out.
I definitely need to work on my pitch more. As for search engines, I tried to target the words Online Diary/Journal (which seem to get searched a lot (1M+ queries per month)) without coming off as spammy.
For privacy concerns you could store the keys (i.e. sleep, running) locally with identifiers to retrieve the values (i.e. hours, kilos) from your database and display/handle them in your app. This way the user keeps track of what exactly is being journaled. Of course, this brings along several other problems like data preservation (local storage).
Sure! I can release it as a library if there's interest. It's basically a go-client querying a redis instance which has been pre-fed with the NLTK categorization for most common words (about 8000 or so).
I think there would be some interest in both the NLP as well as the tracking in general (many say they want to track themselves but privacy is an issue)
That said - it would be cool to see something like this open sources. Currently I'm using excel to track most of the stuff I'm interested in, but I know that there must be a better alternative.