My standard check on this sort of tool is to see how it handles the Sam & Max games. Because they're only sold as a bundle , but they show up in your game list as a bunch of episodes. The general failure mode of the tools is to think that each episode costs the full bundle price.
This tool, indeed, does that. 
(It doing that lets me feel slightly better about my $7k account valuation.)
I've manually updated Sam & Max in the next revision of the database. The tricky thing is, how do you divide a bundle/subscription cost among existing games in that bundle and an unknown / unreleased number of games?
Another interesting fringe case: Half-Life 2: Lost Coast is now only available via a $40 bundle, and is listed under that cost, but is actually just a pack-in bonus to the various Half-Life 2 bundles. How do I represent that? I originally had the $40 listing but updated it to "-" to reflect that the pricing isn't available.
OP and developer of Steam Gauge here - this is my first python project, and I welcome any suggestions / criticisms / advice. If you don't have a Steam account but still want to check it out, feel free to take a look at my profile: http://mysteamgauge.com/search?username=monkeyagent
Here's the summary I posted over on the Steam forums:
What it is:
Steam Gauge will let you get a summary of all the games/apps/etc listed for your Steam account via the Steam Web API. It returns information about those games including hard drive space requirements and cost (base store value, not what you actually paid).
-The list that is constructed allows you to individually select/de-select apps.
-There is a graphical representation of collection percentage selected vs total.
-There are pre-built options to select only a given app type (game, app, movie, etc).
-The app table allows you to sort the list by title, id number, hard drive space, type, and cost (just click the headers).
-If you don't have a vanity username set up, you can use your steamID64 number instead.
-If game information is missing from the database, it automatically adds that id to the list of apps to update in the next database refresh (simply using Steam Gauge helps make it better).
-Finally, you can export the results of Steam Gauge to a csv file, so you can manipulate the data however you wish.
-For now, your profile needs to be public.
-If you use a relatively modern browser, the whole experience is going to look/work better.
-The hard drive values are based primarily on the published requirements (where available) and is otherwise based on manual data checks of game install size.
-The database still isn't complete, especially for new games, and weird exceptions. It should still be helpful for general estimates, but the precision is limited to the data available.
-The mobile layout still needs a little work but should function properly.
-The "Value" column represents the base cost of the game, not what an account owner paid.
-Add more game information including genres, controller support and multi-player capabilities
-Add submission flag for games with missing/incorrect data to help crowdsource what I can't scrape
It has a global rate limiter AFAIK, so it's hit and miss depending on how much load it's under.
Edit: I was trying to datamine Steam several months ago as well, but the amount of data I needed really was unrealistic for this very reason, and it would have just been bad etiquette. I imagine by having this API they're having to deal with a ridiculous number of dataminers. And that's on top of their crazy number of normal users, which I'm sure the rate limiter is there to protect.
I once wanted to calculate how much I have actually spent on steam. The information is only accessible once you log in. So, I wrote a tiny script in js that you can paste in your browser window to calculate that amount. Here it is:
Steam gauge tells me the actual value of my games is $2019.05, but the amount I have spent is $432.75
This doesn't work for me but maybe I'm not using the correct steamID...I've noticed that my "login" name is different than what appears in the top right corner of my steam client. Also, your service may be overloaded without it saying so.
In any case, I looked at my list of games and play time on the Steam Client (on Mac) for the first time. It doesn't appear to be sorting correctly at all...for example, when sorting by "Playtime", the top listed game has "48.4 hrs on record" and then the next game is listed as "0.7"...and then the next game is "69.9" hours...It's kind of astounding to me how Steam is such a wonky, buggy platform given Valve's expertise.
Also, I really doubt that I've played Portal 2 for 70 hours. I guess it counts the time that I've left it minimized on the desktop. Though I admit, I am stuck on one of the latter puzzles and haven't gotten back to it...
A quick note on missing data - I originally scraped data automatically from the Steam store page when a game wasn't found in the database. But I didn't want to add junk data to the database if the scraped data was bad and still need to figure out how to handle those kinds of exceptions. So for now, the tool makes a note of game IDs that people requested but weren't in the database. That way and I can scrape those separately and review the data before I import it to the database. So even though it doesn't auto-populate, your missing entries will likely be updated in the next database refresh (about once a week, depending on how much time I have available).
I've seen a couple of these before, but this is by far the nicest. Love that it spits out a single GB number so you know how big of a SSD you'd need for it all...
Saw a couple of issues you could improve if you like:
1) Civilization 4 (and all DLC) shows up twice, but with no price
2) GTA3/VC/SA all show up twice
3) GTA1+2 show up as $50; you actually can't buy them on Steam, they're just in the $50 GTA Collection
4) Maybe remove 'the' when sorting