teh hayley

I'm hoping today to get a jump on my 1,667 words, and since I feel that my November todo list is incomplete, I figured I'd expand on it here to bump up the word quota.

So to recap, my 2 official goals for November are:

What I didn't document yet is my project list. Specific todos. Things I want to build and/or finish.

So here goes...

Cyclocane

The South Pacific and South Indian ocean tropical cyclone seasons have begun. I'm currently parsing data from the JTWC, but I'd like to get some official data from the individual offices. Of course, this one has to wait until there's an active tropical cyclone, because I currently have no clue where the data is kept for any of the individual RSMCs for either of these basins.

Currently, Cyclocane has storm tracker pages in English and Spanish. I'd like to build German and French versions as well. Based on my traffic stats, French would be more important (and it also happens to be the language of the RSMC for the southwest Indian Ocean), but I actually speak some German, so German would be way easier.

wickedwx

WickedWx is currently my most expensive site to host yet it's actually one of my least traffic'd. If Middleman had existed at the time (or maybe it was that I didn't know about it yet), I would've built this site with Middleman.

It shares a lot of similarities with Cyclocane: semi-frequently changing data, BUT changes that are always server-driven, never user-driven. This basically makes it a perfect candidate for a Middleman rewrite.

This is both high and low on my list at the moment. High because it bothers me that I'm basically wasting money on it when maybe after a weekend or week's worth of work, I could probably cut its hosting bill down to about 20 cents/month.

But it's also low because of

  1. how little traffic it gets
  2. the fact that it's winter so the site ends up being mostly irrelevant until say February-March (though winters are seldom completely quiet because of Dixie Alley (year-round potential for tornadoes))
  3. financially, there's only so far to go with this one. The rewrite drops my hosting bill and then that's it. The rewrite doesn't magically turn this into a money making site, it just makes it less costly to run (point being, it's more important to invest time into the things that have the potential to grow and grow).

national weather alerts

While we're on the topic of weather sites, here's another one I'd like to do.

I'm trying to cultivate this attitude of just-freaking-do-the-MVP-already and this would be a good candidate.

To explain the attitude briefly, the point is that, if it's not going to be a huge time investment to whip together an MVP, then just do it. Then you immediately get some sort of feedback on the project and its potential (and lack of feedback actually is feedback).

Everybody and their brother has weather alerts (and often for free), but I have several sites (and site ideas) that could use some national weather alert data.

Wickedwx was why I started looking at the NWS CAP alerts (I discovered that wickedwx's source of tornado and severe thunderstorm warnings sometimes is out of date/missing data, but it was Cyclocane that actually got me to put together a sort of MVP.

Thanks to the handoff fiasco between the NHC and the HPC, I actually got a jump on building this.

And I do consider it a fiasco because basically all of the coast should've been under Hurricane Warnings not High Wind Warnings. This led to unnecessary confusion that Sandy was going to weaken when in actuality, Sandy wasn't going to lose strength at all (and at points was even predicted to gain strength), it's just that Sandy was going to turn into a different type of storm system before landfall.

For the record, Sandy did become post-tropical before landfall. As of the 7PM NHC update, Sandy was deemed post-tropical. An hour later, she made landfall in New Jersey. An hour. To me, this whole thing reeked of a decision based on technicalities rather than something that was in the public's best interest.

So yeah, rant aside, probably 90% of the CAP alert parser is done. Right now, it's scraping each site individually because the library I found does not support determining what state the alert came from. While I figured this was fine for Sandy (since it's only parsing the northeast states, I definitely wouldn't want to do this for a US-wide site.

The other big thing that's missing is geographic data. Save for certain warnings that have polygon data, most of them are based on the FIPS codes for counties, and I have yet to find out how I would turn a FIPS code into a geographical border. But frankly, this is something that doesn't have to exist for the MVP if I'm just trying to get something out there into the world.

Hack Spanish

I have a lot of sites waiting on me to develop better JavaScript skills. This is one of them.

I decided to write the site during Rails Rumble because I had been stalling on getting anything going and I figured a 48 hour sprint would be the perfect remedy.

The result has been disappointing. Zero search traffic. I made a page for the spanish body parts quiz. No search traffic. I wrote a some additional code so that it would spit out what words were in the quiz so that Google would have a little more to look at. Nada.

At this point, it's basically looking like I'm going to have to write a shortish article on the body parts page just to get google to realize the thing is relevant. To me, this just feels wrong, because I have no idea how to make it not feel spammy. And of course, then that means I have to do something similar for every single quiz I create, which makes it a much bigger time investment.

JS-wise, there's a whole lot more I want to do with the site. To be frank, I've found the multiple choice quizzes on the verb tenses to be worthless. Once you get familiar with the verb endings, the multiple choice answers are a piece of cake. Granted, there are a few things that I could do to attempt to make the quizzes more difficult, but fill-in-the-blank was always the intention. Unfortunately, slickquiz only supports multiple choice and I knew my JS skills weren't going to be up to snuff in time for the Rumble to do something from scratch.

So writing my own quiz system is definitely a todo item. I'd also like to have a completely empty conjugation chart that you have to fill in yourself. Now that would be a challenge.

I also want to have some printable vocab sheets that basically allow you to quiz yourself on paper (it's hard to explain without showing it, and only one foreign language teacher I know ever used this method, so I have to assume it's not all that common).

I also want to do things like a gender quiz, which for Spanish would be pretty easy. So that leads to...

Hack German and Hack French. I can't speak for French (don't speak it yet), but a der/die/das quiz would definitely be more challenging than a Spanish el/la quiz.

What kinda kills me is that, once upon a time, I actually did build a German quiz site. Mostly to help myself out, but my classmates were using it too. I really should've tried to take it public. One of the reasons I didn't I'm sure was the fact that I had written it very insecurely. I'm sure one malicious command could've taken out the entire database because there was no SQL injection projection (one more reason why I love building sites in Middleman).

But really, it was probably more so that I have a habit of building stuff and never properly sharing it. I still haven't learned how to deal with haters on the Internet. Case in point, I cringe every time I see that I've received a new youtube comment, because frankly, the negative comments seem to greatly outnumber the positives. This bothers me so much that I actually have all of the youtube comment emails (for a video I do daily) going to a completely separate email account that I never check. And then I make sure that all comments have to be approved, because it feels a bit embarrassing to have a comment up on my video that basically says I should kill myself.

And it's kind of amazing to me, all of the hate on the internet. I mean, if I don't like a video, I just move on. But the youtube way seems to be to leave a really nasty comment instead. WTF.

Anyways.

green smoothie calculator

I've had this idea in the back of my mind for awhile. Personally, I wouldn't have a use for it at the moment (I'm currently keto so I can't do fruit... and I've never tracked calories anyway), but I thought this is something that a lot of people might find interesting.

The basic idea was to a compile a list of all of the possible ingredients for a green smoothie and then let people pick and choose what they wanted to put in their smoothie, and then they'd be able to see the calories and macronutrient breakdown of their smoothie.

Because I'm all about Middleman and off-loading work to the client, the intention was to always do it in JS. And my JS skills are still not up to snuff to do it (and the other yak shaving bit is that I don't have the raw nutrition data either).

BMI calculator

I've lost a decent amount of fat on keto. Though I think BMI is misguided (especially for those people who don't realize it's nothing but a fancy height/weight chart), I was still curious to see where BMI currently places me.

Though I'd probably eventually find what I was looking for, if I kept looking, I was disappointed by the quality of the BMI calculators out there.

Here's what I'm thinking:

Simple. But here's where the others really disappointed. I shouldn't have to randomly input numbers to figure out how much weight I'd need to lose or gain to get into a different range. Just freaking show me it. And while we're at it, show me the values for similar heights in case I'm borderline between two heights.

I couldn't believe that these calculators didn't have these features. Either you'd be left to randomly punch in numbers, or you were given a visual chart that made it impossible to easily tell where you were (again, don't make me think, just show me the freaking data).

As I said, I'm sure this exists, but I figured this could be an awesome experiment for learning some JS and hey, maybe it would actually make some money if it magically got traffic (since all traffic is still magic to me).

hits worth turking for scraper

This one has been on the list for awhile and frankly I think I can put something together quickly enough that it becomes a "why not".

For some background, hits worth turking for is an online community on Reddit where users find worth HITs of your time (most of the stuff on mturk is slave wages... though with this, you're probably gonna be lucky to make minimum wage, but it's still better than what you'd make without the HWTF community).

Though mturk would technically count as money made online, I wouldn't count it towards the $250 goal because it's completely unsustainable. That said, I am planning on doing more mturk stuff this month anyway so I could actually really use this scraper for myself.

I know that someone has already built one (can't find the address though), but I remember two things. First, I think it wasn't always up, and more important, it didn't have all of the features I'd like. My biggest annoyance with the hits worth turking for, is that you have to sit through adf.ly ads for every single HIT often to find that the HIT is already "dead". Though I think it's good that participants are getting some compensation for helping out the community, it's annoying to have to waste time on dead HITs.

The reddit thing does allow HITs to be marked as dead, but most of the time, it doesn't appear to be getting used (I'm not sure whose responsibility it is to mark the HIT as dead).

I'm not sure how easy it would be to scrape the data, but often times the comments will talk about a HIT being dead (without the link itself being marked appropriately), so it would be a lot nicer if I could figure out how to integrate that data into the scraper.

If I wanted to go overboard, I'd always figure out how to do socket.io and desktop notifications so that someone could just sit on the site and be automagically notified of new HITs.

perfect pitch test

I want to develop perfect pitch. I have this theory that pitches can be memorized (I've already experienced some of this in relation to piano pitches, where I'll recognize a single note in isolation as being from X song). However, I tend to think that if this is true, you'd have to separately memorize each individual instrument. And with sine waves... forget about it. My "method" seems to revolve around memorizing the precise combinations of frequencies that make up a single pitch.

And you know what seems to be missing from all (or most anyway) perfect pitch tests? The ability to select different instruments. Piano seems to be the big one. Everything else gets ignored.

The most popular perfect pitch quiz (based on google results) is extremely limited. I think I can do better.

In fact, I have. But you know what stopped me? The name. I had actually come up with a good domain name, even put it on my monthly todo list to register the thing, and then like 6 days later someone registered it. 6 days!?! And without a good name (and the fact that I was trying to figure out how I would handle creating a paid option), I got stuck and never finished it.

The good news / bad news is that with google's late September EMD update, I feel like I'd have an equal of getting the thing ranked with just putting it on my own site rather than having to give it its own name.

And of course, this project could always use some better JS skills.

IMDB scraper

Speaking of domains, this one is directly related to that.

I've built an IMDB scraper that I'm using on a couple of TV sites, and it would be super easy to expand, but it was requiring that I have a separate domain name for each TV show (well, "requiring" probably isn't quite true, but still).

I'd like to come up with a single domain name I could use, and then just build sub pages off of that site. So again, I'm stuck because of a name. And given its nature, unlike what I said about the perfect pitch thing, I'm definitely not going to put it on this site, because it would be a huge mess.

And there's the thing about domains, $10 at a time is pretty cheap, but when you've got as many experiments running as I do, it adds up.

This IMDB thing also happens to be one of those adsense projects where you make up for things in volume (abysmal CTRs) and since I haven't found a way to get these sites to rank consistently, they all tend to balance each other out, such that with domain costs, they're just breaking even.

So imagine if I could figure out how to build this project around a single $10 domain?

Google Chrome Store app

And finally I think is the todo item without a specific project attached. I'd like to figure out how to get something onto the Google Chrome App Store.

I really think the $250 Chromebook has the potential to be a big hit, and given that I'm passionate about HTML5 type things right now, pursuing the Chrome App Store seems to be a really good idea.

And who knows, maybe it would even be a decent source of traffic.

try low carb

Can't believe I forgot this one, so this will just be brief.

I've been wanting to have a place where I could store keto / low carb recipes I wanted to try / have tried.

I've also got insane plans about some of the useful features I could add to elevate it from just a simple recipe site. But for now, I'll just be happy to have one place to have this information stored (instead of the current scraps of paper that recipes get written on).

Not only that, it has the potential to contribute towards the November $250 goal. So I should totally get cracking on it.

It actually exists as of last night in MVP form. But the data structure is still a mess. I'm trying to put stuff into YAML frontmatter on individual pages, but the Middleman instructions weren't helping me (perhaps some of the method names were out of date), and I never could figure out how to properly access the individual frontmatter from other pages. Plus, for all of the plans I have for it, it's probably going to be a nightmare for it to be stored in a whole bunch of different files, so I'm getting stuck because I'm thinking too much about how it needs to be structured in the future.

So I think that's finally it.

earthquake sites

Totally wasn't the end of the list. I completely forgot about the Earthquakes Today spinoff.

So Earthquakes Today was actually the first place where I noticed the late September Google EMD algorithm update.

I went from around 100 visitors/day to basically only the people who had bookmarked the site.

The site has basically become unfindable. Wanna hear something disgusting? I'm using the phrase "was there an earthquake today? you bet" and I know that one of my users finds my site by "earthquake today you bet".

A f*(&ing spam site outranks me for my own phrase.

I mean, I can understand that I no longer rank for earthquakes today... but now I don't even rank for my own unique phrase.

So ET is a good example of why brandable domains are a good thing.

In the aftermath, I decided that this was the push I needed to rename the site (I was never really comfortable with the fact that I had a site that was a .org).

I've written about it before (though I'm not sure I've posted that article on this site yet), but I came up with two names:

I thought this was the perfect opportunity to actually test the names to try to gauge which one would be more popular.

So over September, I ended up running 2 tests.

First, I ran a test on Cyclocane using Optimizely to test how many users clicked on earthquake storm vs seismic storm. The result? Inconclusive. One was leading by a little, but the last time I looked, the two were basically neck and neck (and at this point, my free trial is over so there will be no more testing on that one).

Second, I posted a note on ET itself and asked users to click on the name they preferred. I did the poor man's A/B test by putting Google Analytics on both sites and then monitored which site got the most traffic. Same result. No clear winner.

At this point, I'm thinking the winner is going to be determined by which one looks like I can actually rank for it.

Seismic Storm is ranking with its bare "coming soon" page. "Seismic Storm" was once ranking about the same place with its coming soon page and then fell. So I decided I'd "port" the ET page over to it but then base it around 7 day data. It's never recovered.

Now, the problem is that maybe Earthquake Storm was too similar to Earthquakes Today's content, so google has marked it as a copycat site. I have no idea.

Meanwhile Seismic Storm is still ranking with its tiny page.

So basically, Seismic Storm may win out simply because I'll be more easily able to rank for my own site's name.

So my plans overall are to "spin-off" what I've done with Earthquakes Today (and basically leave ET alone... because, hey, it may one day recover its rankings) and create a new site. I want to put a page for "was there an earthquake today" that detects your location and then shows you your last earthquake (and probably recent earthquakes as well).

I've got other things planned, but of course, until I can figure out what name I really want to use... I just can't work on it, now can I?

card games

I find it amusing that I forgot this one, because this is debatably my highest priority this month.

Speaking of stealing ideas (you'll read about it in the next section... unless I've forgotten yet another November project and then it will be a little further down), this plan grew directly out of a stolen idea.

I read about someone who had created a few card game sites and he was making a non-trivial amount of money on Adsense. I emailed some friends about it because I thought that was fascinating, especially since these were basically set it and forget it type sites. Build it once, it's totally static, and then just watch it take off (or in my case, it would be more so just watch it get zero traffic).

So I kept that idea on the backburner for awhile. I don't remember what the tipping point was, but recently-ish I decided that this could really be the thing that finally elevates me above just earning beer money online. Plus, I thought I had a potential leg up on the guy that I stole the idea from: I'm fascinated by languages. I can come up with one game and then translate it (or have it translated) into many different languages, and thus hopefully increase the traffic.

Speaking of which, to be clear, I believe that putting ads on game sites is one of those "we'll make up for it in volume" situations: most people won't click, so you're hoping to get enough traffic to make your abysmal CTR worth it.

So I did some keyword research on spanish card games (I still don't really speak Spanish, but figured I'd cross that bridge when I came to it). Eventually, I discovered a decently high traffic keyword where the EMD was still available.

Granted, the late September update makes the exact match domain less worth it, but I liked it because it was fairly short and language neutral (for instance, compare "Spider Solitaire" (English) and "Solitario Spider" (espaƱol) - this domain is a decent compromise).

Based on Google Trends' lack of country data, I also had no idea what countries were searching for this specific term, so my thought was that I'd register the domain, put analytics on it, and then hope it got enough type-in traffic that I'd be able to draw conclusions about which languages I would need to support.

So why haven't I just said what the domain is? I'm embarrassed by it. On the subject of stealing, as part of the experiment, I decided to basically rip off this guy's solitaire code to get the site bootstrapped. I absolutely want to rewrite it. I think it's dirty what I've done and I don't feel great about it, but I felt it was more important to start getting some raw data on the site and worry about a rewrite later.

So my big priority is to learn enough JS to be able to build this sucker on my own and then I can feel free to promote it without guilt. Not that I've learned anything about promotion though.

steal my ideas??

So I realize that I've put a lot of ideas out there. Supposedly though, you have to nag people to steal your ideas, so I shouldn't be worried.

And besides, once an MVP is out in the world, your idea is now vulnerable to being stolen anyway.

And plus, a lot of my ideas basically come from stealing other people's ideas, so who am I to think that I should magically be the exception?

For me though, I think it's all about accountability. I could make the argument that making the list public now gives me a reason to build an MVP before someone beats me to the punch.

More so, I like that I now have a single list, in a single place, that reminds me of what I think is important to accomplish this month.

In reality though, my track record means that I'll be lucky to get one of these things done.


teh hayley