Congratulations on your new data portal!
You’ve made a difficult decision, but Data Portals International Ltd would like to assure you that you’ve made the right choice. This quickstart guide is meant to help you get on your feet with your new portal as effectively as possible, as we know it is difficult to be an early adopter of new technologies.
Step 1: Acquire Your Data
For your new data portal to function properly, you must acquire data to load into it. The more useful the data that goes into it, the more useful the results coming back out will be. Here are some easy steps for identifying and maintaining your data:
1a. Pick the right data
First, you must select the right data to put into your platform. It should be meaningful to the eventual users of the portal. People who answer public inquiries are fantastic resources to help you to learn what your constituents are interested in, as are librarians. When in doubt, take a look at what data other jurisdictions have put out that are most popular, particularly jurisdictions that are similar to your own. Sending out polls on social networks can be very successful, too.
Try to avoid data that is already publicly available, like the basic demographic data the Census provides, or data that is summarized at too high a level to be useful, like the number of wizarding schools in your jurisdiction.
Aim for quality, not quantity—if the first few things people wander across aren’t assembled with care and an eye towards eventual use, they will assume forever that everything on the site is useless.
1b. Make it friendly
The next step is to make sure the data is friendly to both humans and robots. While this might seem at first to be an insurmountable diplomatic challenge in the face of the history of all of literature and cinema—relax, and take a breath. It turns out that the machines you care about are built by humans, and they end up having the same needs:
The first key to satisfying your demographic is to ensure the data is structured. Avoid formatting tables for presentation—section summary lines and multiple-level headers may make a lot of sense for a paper publication, but they confuse people trying to sum up an entire column. Make sure numbers are numbers and letters are letters.
The second key is making sure the data is clean and understandable. Nobody knows what violation code 1138F is, but everyone understands “maggots found in refrigerator,” even the robots. The robots, however, start to get annoyed when it turns out that Bob over in Records couldn’t spell “maggots” and half the instances of 1138F actually read “magots found in refrigerator.”
1c. Automate it!
Unfortunately, you own the DP-3000N, and not the DP-3100N, and so you don’t have access to our telepathic instant data ingest system (call our sales staff today to upgrade!), so you’ll have to rely on “real” methods like programmatic updating. But, automating early is definitely the correct approach. Getting the data to update by itself will make it immensely more useful to everybody in the long term.
Step 2: Present Your Data
Now that you have a library of amazing, friendly, automated data to work with,
the next step is to get people to it.
This means some information architecture and design, but fret not—follow these three simple steps and you’ll be on the right path.
2a. Teach people something immediately
Nobody cares about the number of datasets you have, or how excited their elected officials are about the opportunities created by this newfangled data portal thing. People do care about what’s going on around them—so show them that! The number of datasets may not be interesting, but the number of biblical miracles performed in the last year in their district is certainly remarkable to most people.
Teaching people a useful, salient fact immediately informs them this site might contain other information useful to them, and encourages them to stay and browse the data.
2b. Show something live
The most popular pieces of data are always the ones that are the most live. No need to seek out Dr. Frankenstein—if you already followed step 1c, you’re already set. Self-updating data, particularly data that updates on an interval of minutes, is necessarily interesting. The City of Seattle’s most popular data are its near-real-time view of the fire-related 911 calls coming in. Put this sort of data front and center outside the confines of a dataset, and people will be immediately hooked.
2c. Provide great drill-downs into your data
Now that people understand that your fully assembled site contains some awesome information, it’s time to make sure they can find their way to all the data they care about. The key is to take the things people are interested in and provide a view of the world that breaks down in similar ways. How to do this varies a bit—topic-based organization is always a good start, though if possible you should repeat steps 2a and 2b for each topic drilldown.
Step 3: rinse and repeat
Now that you’ve gotten through the majority of the work, it’s time to rinse and repeat.
No, we’re serious
First, lift the center flap on the DP-3000N, and give the inside a good rinse. We’ve found that a garden hose works pretty well. Do make sure you unplug it before you do this; we cannot be held responsible for any shocking effects that may result otherwise.
Rinsing is the easy part—now you must repeat. Check all the answers you came up with to the above questions, and make sure the content you’re releasing is what people want to see. Watch ordinary human beings use the site, and resist the urge to assist them: where they struggle is your chance to improve the organization and presentation.
Nothing will come out of your data portal if you don’t seek to constantly improve upon it. A good data portal makes for happy robots, humans, and puppies.