Data Analytics

I have a bit of a love/hate relationship with data analytics. On the one hand, I think there's a lot of value to be drawn out of analyzing data and looking for trends and patterns to assist in decision making. On the other hand, I think the process of analyzing the data is tedious and time consuming. I don't like crunching and poring over all the data. I'm all for collecting as many data points as possible but ideally I would have someone that analyzes the data for me and finds all the useful nuggets of knowledge.

I bring all this up because lately I've been doing a lot of data analysis in and out of work, and unfortunately right now I'm not in a position to have others do the analysis for me. So that means I've been crunching a lot of data lately. This actually hasn't been too bad because I've been focused on the results and the payoff that might come out of the data.

At work I have two data analytics tasks on my plate. Now normally I would not be too happy if my boss put me to work crunching data, but both of these tasks have come from my own initiatives and are part of the 'innovations' task force that I'm leading. Unfortunately the innovations budget only covers me right now so I have to provide the hard data to support my initiatives on my own and that means crunching rows of data. One of the tasks is to prepare quoting sales data that our system has been collecting for years and use this to improve our consumer profiling process. The other task is to analyze data from a project tracking system we've been using for the past year and look for opportunities to improve our project portfolio management process. I've found some solid actionable areas of improvement here and I'm now preparing a write-up to present my recommendations to management. (I'm actually using a lot of the ideas and concepts from my Technology Management class for this initiative. This whole MBA thing is paying off already!)

Outside of work I've been playing around with Twitter data and digging into the area of content curation. I've tested a lot of the Twitter clients out there and I haven't really found anything I like that presents my Twitter feed to me in a way that is manageable and still enables discovery of timely and relevant content. So I've been making my own Twitter curator that tries to organize my Twitter feed and shows me exactly what I want to see based on several different data points. Some of the points include friend quality, follower quality, post frequency, list memberships and link quality. It's not production ready by a long shot (and it may never be), but I'll try to post screens of my Twitter curator sometime in the near future.

Music moods

Just last week I wrote about a new feature I'd like Pandora to add (music maps) and this week I've read about some new music sites that have this and other cool features included. I stumbled upon these sites through an article in Mashable called "5 great ways to find music that suits your mood".

A couple sites, like Musicovery, have the music map feature, but more interestingly all the sites listed try to match the music to your current mood. I think this is a great idea that would suit me really well. I have my 'Rock' station, and my 'Hip-Hop' station, but sometimes I want my Rock & Hip-Hop to match the mood I'm in right now. Am I in a mellow rocky mood, or a hard party rock mood? Do I want Hip-Hop that I can nod my head to, or do I want Hip-Hop to get me pumped and ready for the night? Sure, I could make Pandora stations for each mood, but overall I think the artists (and even albums) would overlap. It's really the individual songs and the feelings they invoke that would be driven by the "mood" settings.

I'll probably give one or two of these sites a spin and see how well they categorize "mood" at the song level. I wonder how their music catalog selection holds up against Pandora and the other bigger online players. Really I just want Pandora to add this feature so I don't have to keep track of multiple music sites and can keep all my musical preferences in one spot.

Evernote users, let's hear from you

"Evernote" and "slow" continue to be the top Google search keywords leading to my blog. My site is actually one of the first ones that come up when you Google "Evernote slow". I find this a little funny considering I've only written about Evernote 3-4 times, but I guess I've struck a chord near and dear to frustrated Evernote users.

I say 'frustrated' because overall I love Evernote, I just wish it was a bit more responsive at times. The Windows client continues to frustrate me with its constant hanging and memory hogging; so much so that I avoid opening it unless I really need to do extensive organizing or browsing of my notes. If I want to do a quick search of my notes I find it faster to jump onto the website and avoid the slow load times of the Windows client.

Also, whenever I want to clip web content to my notes I always try to use Firefox and the Evernote plug-in because it's 10x better than the Internet Explorer version. Why is that? Well, for one, the Firefox client doesn't load up the Evernote client, instead it provides with you with a quick and easy interface to add in the content to your journal. The IE plug-in sends the content to your Evernote client, which then takes a couple of seconds to respond, and then takes a couple more seconds to load up the Evernote client so you can annotate the web content. Big FAIL for the IE Evernote plug-in.

Alright, so those are my frustrations with Evernote and slow. What about everyone else? Feel free to share your experiences in the comments below. Who knows, if an Evernote rep Googles "Evernote slow" and stumbles upon my little blog hopefully we can provide this person with detailed experiences and suggestions that they can use to improve their product.

Pandora music map

I've been fiddling around with my Pandora stations a bit lately, marking songs that I like and dislike as I listen throughout the day at work. Unfortunately this caused my favorite stations to destabilize a bit and I started to hear a lot of music that I didn't want to hear or that didn't belong in that particular station. I finally had to resort to going into the Pandora control panel and 'rolling back' all of the likes and dislikes I've given recently to get my stations back in order.

This mini fiasco led me to wonder about Pandora's algorithm again. I'm not quite sure what action I took that caused some of the 'wrong' music to leak into my stations but I really wish I had visibility to the whole process behind Pandora. I've said before that I like the simple thumbs up/down model but I wish we had more advanced features available to us. I understand Pandora probably doesn't want to reveal its algorithm to the public but I think there are certain features it can add that would make Pandora better and easier to use.

One idea I like is adding a 'music map' view in Pandora. It would be pretty cool to see the music map representation of your stations and to see what sort of artists are at the core and fringe of your musical preferences. Maybe you want to hear a little bit more from a fringe artist, or a little bit less from a core artist? You could just drag the nodes of each artist in/out and have your station reflect that change. It would essentially be a visual representation of your music stations.



Another cool thing you could do with the music map feature is to see what effect an action you take will have on your station's feed. For example, when you're about to like or dislike a song, it would be pretty interesting to see on the music map what new nodes appear or what existing nodes move around. That way you can get a better feel for the composition of your station and what influence your actions are taking on it. This feature sure would have saved me a lot of hassle lately!

This new feature would also be really cool on the Pandora iPad app. The iPad's large interactive touch screen makes it ideal for playing around with a music map, and the feature would keep Pandora users more actively engaged while they're listening to their music. I think it would be a great addition to the "ultimate lean-in Pandora experience".