VOTING WITH YOUR TWEET: An experiment in political forecasting

This post was originally published by California News Service and was written by Mark Huberty.

Now that the election’s over, we can say a few things about how we did using Twitter to predict the winners.

First, we didn’t do so well. Using our predictions for what share of the vote Democratic candidates would earn, we predicted the winner about 75% of the time. That sounds good–after all each race has only 2 candidates, so maybe doing better than 50% accuracy isn’t so bad. But knowing races are usually biased towards the incumbent would have allowed us to predict correctly about 85% of the time.

Continue reading →

VOTING WITH YOUR TWEET: An experiment in political forecasting

The Voting with Your Tweet project is a window into an emerging field in political research — the automated, computational mining of personal data. At this stage it is very much a work in progress – a draft for comment if you will, in the tradition of all serious academic research.

So why publish intermediate results on a news site?

Because the mining of social media data is becoming increasingly important in politics. Our goal is to help journalists and the public better understand new political prediction techniques and to help them become critical consumers of this kind of research.

Continue reading →

The following chart was made using freeDive and Goggle Finance functions in a Google Spreadsheet. It updates automatically throughout the day. There can be a delay of up to 20 minutes.
Continue reading →

My bed head has been a daily source of amusement for my wife since we’ve been together. So last year I took a picture of myself as soon as I woke every morning and put it together in the video above. While I did miss a few days due to travel/camping trips, every bit of my hair is real, unadulterated and uncoiffed.

I have to say I wasn’t sure if this project would come together. Putting it in a video format that I liked took longer than I thought it would. But in the end it was worth the effort. I hope it makes you giggle at least a little bit

Special thanks to my wife Magaly, who put up with this and suggested the “hey ladies” image (the only one not from 2011). The music was sampled and mixed by Girl Talk (AKA Gregg Gillis) on his All Day release. Its use is allowed under a Creative Commons license.

Got my 1977 VW Campmobile back from getting worked on last month and it’s now mechanically and physically sound. Was talking with a vagabond friend and he told me about Boondocking, the art of camping for free. A quick Web search resulted in a database of locations and reports by Andrew Koransky at his site boondocking.org. So I whipped up this map in Google’s Fusion Tables. It’s a little easier to read in the larger version.

Been really busy the past two weeks with KDMC’s Digital Storytelling Workshop. Here’s a look at the types of media the fellows work in. This treemap has been reconfigured to create some visual interest but preserves the area of the shapes.

Just for kicks, I decided to visualize the schedule for KDMC’s Digital Storytelling Workshop using nothing but html and CSS. It’s a little hacky, but that’s how I roll.

It’s fortunate that I can walk to work most days, so my first post is a look at my work commute for the month of November. Thicker lines indicate more frequently traveled routes. The middle route, while not the shortest, is the most scenic.

This site is a place where I can publish random statistical visualizations. I see data in strange and sometimes amusing places (at least to me). I hope to share them on an irregular basis. The links in the sidebar point to what I consider the best of the data blogs. I have also included a feed of my data visualization Twitter list. It includes some really smart people who share great content.

I’ve also decided to share a collection of visualization tutorials that I’ve built over the years. There’s plenty of room for that to grow so if you know of something that should be included, please let me know.

I’m a visual journalist and and instructor at the Knight Digital Media Center at UC Berkeley. I have produced and managed graphic for newspapers, web sites and a TV news station. You can find out more about me at my personal site, lendegroot.com.

This is an HTML 5 site. It will automatically resize to fit your PC, tablet or smart phone. It uses a WordPress child theme that I created and is in a state of constant refinement. I apologize in advance if something doesn’t work. Also, many posts include technologies that require a modern browser (Chrome, Safari, Firefox or IE9).

If you see something you like, please tell me in the comments. I would consider it a coin in this dancing monkey’s cup.