I'm back in Canada, and also making my first post since I was last here. Maybe I have to move to Canada if I'm going to get blogging regularly! I'm at the Big Data and the Cloud 2012 conference in Edmonton. It's at the Fantasyland Hotel, West Edmonton Mall (once the biggest in the world). On ringing to make my reservation I was told "Hi, this is Britney, how can I make your stay with us more fun?" I was very discrete in my response. Just a few steps down the corridor from the hotel is this:
It's been a very worthwhile few days with lots of interesting presentations and conversations, and also some worries, especially for the educational application of analytics. I'll be going into that in future. For now I'll just mention that Dwayne Harapniuk, VP Academic Concordia University Alberta, made a very gung ho presentation about analytics, including references to Davenport, a Harvard analytics guru. According to Davenport it seems you have to use analytics across the organisation because you only get the benefits when you get data from the whole entity. Dwayne says this means everything must be based on analytics, and we all have to accept that we are evidence based in everything we do. This sounds a recipe for ratcheting up managerialism (and deliverology?) still further, and is in conflict with the start small and experiment advice which was given in other sessions. It's all very bottom line inspired: “We are student centred, so we go by student grades”. Student centredness has gone on a long journey since I started teaching!
Jurji Paraszak, of IBM, (who I was largely in sympathy with) talked about the importance of models. But he only talked about them after the data was collected. There was no mention from him or other presenters about the models which lead to the choice or creation of data sources. It seems to me that there is always a theory lurking in the background, informing the data gathering (and the areas where no data is gathered), and it was the lack of an explicit theory was what worried me most about what Harapniuk's talk.
It's been a very worthwhile few days with lots of interesting presentations and conversations, and also some worries, especially for the educational application of analytics. I'll be going into that in future. For now I'll just mention that Dwayne Harapniuk, VP Academic Concordia University Alberta, made a very gung ho presentation about analytics, including references to Davenport, a Harvard analytics guru. According to Davenport it seems you have to use analytics across the organisation because you only get the benefits when you get data from the whole entity. Dwayne says this means everything must be based on analytics, and we all have to accept that we are evidence based in everything we do. This sounds a recipe for ratcheting up managerialism (and deliverology?) still further, and is in conflict with the start small and experiment advice which was given in other sessions. It's all very bottom line inspired: “We are student centred, so we go by student grades”. Student centredness has gone on a long journey since I started teaching!
Jurji Paraszak, of IBM, (who I was largely in sympathy with) talked about the importance of models. But he only talked about them after the data was collected. There was no mention from him or other presenters about the models which lead to the choice or creation of data sources. It seems to me that there is always a theory lurking in the background, informing the data gathering (and the areas where no data is gathered), and it was the lack of an explicit theory was what worried me most about what Harapniuk's talk.
No comments:
Post a Comment