This weeks lecture mentioned recommender systems and I thought I'd share little of my experiences with them and how I use them. Lecture and many of the courses blog posts mention amazon.com as an example of site using recommender systems, but I think I'll talk about bookdepository.com as I've used it more. (And also because it's an excellent internet book store. Cheaper comics than here in Finland and books in english too.)
Often when I'm going to the store I'm looking for next book in a series.I may not remember the exact title, but I do know the name of the previous book. So I type the name of the previous book into the search field and hit search. This is the part where recommender systems come in. Usually what I'm looking for shows up in a box showing "Other people who viewed this bought"-items.
So I use recommender systems as a kind of extention to search. Impulse buying is other way of using them that comes to mind, but I usually leave that to actual stores. Internet stores just don't produce that impulse to buy something on a whim, when you just see the product but can't touch or examine it.
Saturday, February 19, 2011
Thoughts about recommender systems
Lecture slides mentioned amazon.com as one example of site using recommendation system. I have myself found out that amazon's recommendations work extremely well. Only one starting point like book or audio record is needed and you can start to follow recommended links without need to use search or categorized navigation anymore. I have actually ended several times to order one of the recommended books.
Could recommendations be the most important way of browsing data in web site? Spotify and Last.fm (more about Last.fm in Minna's blog) music services has recommendation systems called introducing similar sounding artists to the user. Recommendations are based on data what users are really listening. In this case of music library recommendations are very convenient way to browse data. System allows user to find new bands and music without prior knowledge. It could be tedious to browse library of million records by clicking multilevel categories of different genres. It could be hard to imagine music service without recommendations.
How about implementing recommendations for the web site of daily newspaper with articles of various topics? I think in this case it's not so easy to collect data and give good recommendations. User behaves differently compared to previous examples. I guess that users are mainly interested in newest material. However, some newspaper offers recommendations based on similar keywords or popularity of the article with varying results.
Recommendations may be valuable addition for web site usability. Some cases it can even replace commonly used navigation structures. However this is highly depending about site content, user behaviour and how easy it is to collect meaningfull activity data for the basis of recommendations.
Could recommendations be the most important way of browsing data in web site? Spotify and Last.fm (more about Last.fm in Minna's blog) music services has recommendation systems called introducing similar sounding artists to the user. Recommendations are based on data what users are really listening. In this case of music library recommendations are very convenient way to browse data. System allows user to find new bands and music without prior knowledge. It could be tedious to browse library of million records by clicking multilevel categories of different genres. It could be hard to imagine music service without recommendations.
How about implementing recommendations for the web site of daily newspaper with articles of various topics? I think in this case it's not so easy to collect data and give good recommendations. User behaves differently compared to previous examples. I guess that users are mainly interested in newest material. However, some newspaper offers recommendations based on similar keywords or popularity of the article with varying results.
Recommendations may be valuable addition for web site usability. Some cases it can even replace commonly used navigation structures. However this is highly depending about site content, user behaviour and how easy it is to collect meaningfull activity data for the basis of recommendations.
Thursday, February 17, 2011
Blogging policy
Blogging policy
Groupwarecourse inspired me thinking, that it could be nice to do a little research in political parties in which kind of writing, blogging page they have!
The parliament elections is coming and one would think, that it appear as increase activities in Blog writing.
What are the main parties writing in their Blogs in their home pages?
Are party members acting before the elections? Political parties scribbling agitation increases or not! This should show as an increase in writings and diverse writings and discussion about the operative actions for common goal of parties.
What are Web tools and styles, which parties use. How parties show up values and options in blogs. All these styles vary many ways from each others.
Where blogs are placed in the homepages? What are in these pageses?
Very shortly
SDP http://www.sdp.fi/
Päivän blogi: is in the middle of the home page and it will be truncated on the display actual, not emerging topics and arguments and counter-arguments in special cases. These topics take distance to the actual political issues.
CENTRE http://www.keskusta.fi/
Blogeissa nyt: under the title has been selected topics to carry the party leader withdrawals and her arguments and as well selected comments against these ideas.
Discussion topics are not handling most burning political issues
KOKOOMUS http://www.kokoomus.fi
In the home page I do not find out the real blog. Instead, there is Jyri pages in, with Jyrki describe his thinking as a one voice of the party. No at diffused thinking or ideas are not presented.
GREEN http://www.vihreat.fi/
Vihreä blogi: is in the front page of the main webpage. There are given lot of space to active party members to express their ideas and opinions. Blogs are illustrated with advertise photos of all writers. Outside writers are pulled out from the website.
PERUSSUOMALISET http://www.perussuomalaiset.fi/
In webpage are not real blog. Instead there is the MIELIPITEET page. In “mielipiteet” page are selected writings of party members around Finland. The Timo Soini writes his own “ploki” in the webpage.
KRISTILLISET http://www.kd.fi/KD/www/fi/index.php
KD-blogt: You can find the KD-blog in the home page. There are selected writers with photos expressing their ideas. "Permissible" for discussion topics are proposals to build the good value society.
SFP http://rkp.fi/
Kolumnit: In the homepage there is not blog instead there are collected articles under the “Kolumnit”. There have been collected opinions. In this context does not blog just to talk.
It is obvious after looking these party homepages, that the debate happens in Face Book and WebPages of local party divisions. That's what this mini-review revealed in the first hand.
Another finding, which everyone knows, is that Parties do not discus in their Homepages about real problems.
The unemployment and how it can avoid or the national deficit and how the government should deal with it.
The coming parliament election is in a way or other possible to seen in blogging. The parties avoid discussion of real things. Which everyone in Finland has discovered without looking at websites of political parties.
Erkki
Groupwarecourse inspired me thinking, that it could be nice to do a little research in political parties in which kind of writing, blogging page they have!
The parliament elections is coming and one would think, that it appear as increase activities in Blog writing.
What are the main parties writing in their Blogs in their home pages?
Are party members acting before the elections? Political parties scribbling agitation increases or not! This should show as an increase in writings and diverse writings and discussion about the operative actions for common goal of parties.
What are Web tools and styles, which parties use. How parties show up values and options in blogs. All these styles vary many ways from each others.
Where blogs are placed in the homepages? What are in these pageses?
Very shortly
SDP http://www.sdp.fi/
Päivän blogi: is in the middle of the home page and it will be truncated on the display actual, not emerging topics and arguments and counter-arguments in special cases. These topics take distance to the actual political issues.
CENTRE http://www.keskusta.fi/
Blogeissa nyt: under the title has been selected topics to carry the party leader withdrawals and her arguments and as well selected comments against these ideas.
Discussion topics are not handling most burning political issues
KOKOOMUS http://www.kokoomus.fi
In the home page I do not find out the real blog. Instead, there is Jyri pages in, with Jyrki describe his thinking as a one voice of the party. No at diffused thinking or ideas are not presented.
GREEN http://www.vihreat.fi/
Vihreä blogi: is in the front page of the main webpage. There are given lot of space to active party members to express their ideas and opinions. Blogs are illustrated with advertise photos of all writers. Outside writers are pulled out from the website.
PERUSSUOMALISET http://www.perussuomalaiset.fi/
In webpage are not real blog. Instead there is the MIELIPITEET page. In “mielipiteet” page are selected writings of party members around Finland. The Timo Soini writes his own “ploki” in the webpage.
KRISTILLISET http://www.kd.fi/KD/www/fi/index.php
KD-blogt: You can find the KD-blog in the home page. There are selected writers with photos expressing their ideas. "Permissible" for discussion topics are proposals to build the good value society.
SFP http://rkp.fi/
Kolumnit: In the homepage there is not blog instead there are collected articles under the “Kolumnit”. There have been collected opinions. In this context does not blog just to talk.
It is obvious after looking these party homepages, that the debate happens in Face Book and WebPages of local party divisions. That's what this mini-review revealed in the first hand.
Another finding, which everyone knows, is that Parties do not discus in their Homepages about real problems.
The unemployment and how it can avoid or the national deficit and how the government should deal with it.
The coming parliament election is in a way or other possible to seen in blogging. The parties avoid discussion of real things. Which everyone in Finland has discovered without looking at websites of political parties.
Erkki
Wednesday, February 16, 2011
Need new music?
If that question sounds familiar, Last.fm can be the answer. Last.fm is a nice music recommendation service that uses The Scrobbler by which all music listened is added to user’s profile. Based on the tracks user has played with some media player, Last.fm helps user to find similar music.
Some features of Last.fm which I find useful:
- Neighbours: People listed with similar music taste, and possibility to see their top tracks and artists.
- Own profile and charts: Artists and tracks played.
- Music recommended for user by Last.fm
- Artist pages: The most popular tracks of some artist. It is especially useful with some new artist to try to listen some popular tracks first. Also it is interesting to see, what tracks of some artist the majority of the Last.fm users like most. On the artist page user can also look “other similar artists” to some artist.
There are many other useful features too, like events, tags, loved tracks, radio stations and other information on artists pages, but those features I listed earlier are the ones I use most. In general, Last.fm has offered for me what I have been looking for. I have used it during some years, and with that I have found a lot of new interesting artists and tracks similar to my music taste.
http://www.last.fm/
Some features of Last.fm which I find useful:
- Neighbours: People listed with similar music taste, and possibility to see their top tracks and artists.
- Own profile and charts: Artists and tracks played.
- Music recommended for user by Last.fm
- Artist pages: The most popular tracks of some artist. It is especially useful with some new artist to try to listen some popular tracks first. Also it is interesting to see, what tracks of some artist the majority of the Last.fm users like most. On the artist page user can also look “other similar artists” to some artist.
There are many other useful features too, like events, tags, loved tracks, radio stations and other information on artists pages, but those features I listed earlier are the ones I use most. In general, Last.fm has offered for me what I have been looking for. I have used it during some years, and with that I have found a lot of new interesting artists and tracks similar to my music taste.
http://www.last.fm/
Monday, February 14, 2011
Group identity in MMO's
During the lecture there was some discussion about MMO-games and communities in online spaces. How does one's group identity form in these online societies and how it can be developed?
Yusuf Pisan studied the formation of group identity in guilds (social communities inside an MMO), in World Of Warcraft MMO. He based his research on group-identity theory, which consists of three components:
- Cognitive component, which represents person response to his own identity and how he categorizes himself as a part of a group.
- Affective component, which represents the group's attractiveness from the person's viewpoint and the group's ability to help the person to reach his goals.
- Behavioral component, which represents the group's dependancy on each other and it's ability to coordinate activities towards common goal.
Pisan's research showed that of these three the most important in the formation of a guild's group identity was the affective component, in other words, how much the members of a guild liked each other's company. It should be noted though that all three components were highly dependant on one another, so there wouldn't be a case were affective component was very low but cognitive component was very high, meaning that people wouldn't like each other but still wanted to feel part of the same group.
The importance of the affective component strengthens the idea that virtual worlds act as social spaces in which people socialize with each other just as they would in a coffee shop or a bar. Dmitri Williams, Scott Caplan and Li Xiong found in their research that the use of voice in communication in online spaces had highly positive effects on people's relationships. The relationships grew a lot deeper and the trust between people increased when they moved from text-only communication to voice communication using VoIP-services.
This leads us to the conclusion the online spaces which allow people to use voice communication as a part of everyday communication, would be the ones which also develop the greatest feeling of group identity in the communities.
References:
[Pisan, 2007] Yusuf Pisan, My guild, my people: role of guilds in massively multiplayer
online games. Proceedings of the 4th Australasian conference on Interactive
entertainment, RMIT University, 2007, Article No. 20.
[Williams et al., 2007] Dmitri Williams, Scott Caplan, Li Xiong, Can you her me
now? The impact of voice in an online gaming community. Human Communication
research (Volume 33, Issue 4), International communication association,
2007, 427-449.
Yusuf Pisan studied the formation of group identity in guilds (social communities inside an MMO), in World Of Warcraft MMO. He based his research on group-identity theory, which consists of three components:
- Cognitive component, which represents person response to his own identity and how he categorizes himself as a part of a group.
- Affective component, which represents the group's attractiveness from the person's viewpoint and the group's ability to help the person to reach his goals.
- Behavioral component, which represents the group's dependancy on each other and it's ability to coordinate activities towards common goal.
Pisan's research showed that of these three the most important in the formation of a guild's group identity was the affective component, in other words, how much the members of a guild liked each other's company. It should be noted though that all three components were highly dependant on one another, so there wouldn't be a case were affective component was very low but cognitive component was very high, meaning that people wouldn't like each other but still wanted to feel part of the same group.
The importance of the affective component strengthens the idea that virtual worlds act as social spaces in which people socialize with each other just as they would in a coffee shop or a bar. Dmitri Williams, Scott Caplan and Li Xiong found in their research that the use of voice in communication in online spaces had highly positive effects on people's relationships. The relationships grew a lot deeper and the trust between people increased when they moved from text-only communication to voice communication using VoIP-services.
This leads us to the conclusion the online spaces which allow people to use voice communication as a part of everyday communication, would be the ones which also develop the greatest feeling of group identity in the communities.
References:
[Pisan, 2007] Yusuf Pisan, My guild, my people: role of guilds in massively multiplayer
online games. Proceedings of the 4th Australasian conference on Interactive
entertainment, RMIT University, 2007, Article No. 20.
[Williams et al., 2007] Dmitri Williams, Scott Caplan, Li Xiong, Can you her me
now? The impact of voice in an online gaming community. Human Communication
research (Volume 33, Issue 4), International communication association,
2007, 427-449.
Saturday, February 12, 2011
Thoughts about GSS
In the last lecture there was talk about Group Support systems and how some of them even need separate facilitators, who before meeting plan the meeting and what tools to use and during the meeting uses the tools. This whole facilitator-idea made me wonder whether all the fancy GSS really help the meeting. I mean, if you need a special trained professional just to run them, wouldn't it just be easier to hold the meeting in a old fashioned way?
Of course, I can see the benefits of - for example - computer counting votes or other manual things that would take plenty of time if conducted manually. But if I imagine a system that would do that or other similar simple tasks, it hardly seems like a system that would need a facilitator to run it. Of course, this is all just thoughts based on how I imagine GSS's work. I don't have any personal experience with GSS's, and I don't really know what kind of tools they provide or how hard they are to use.
Other thing that raised a thought was study by De Vreede et al.[1] where they studied causes for GSS meeting failures. Most of the failures had something to do with problems with goals and far less common were the ones with problems with technology. So most of the problems seem originate from the participants, not the tools. Just like in Jami's post here, it seems that it all comes to the people using the tools.
References:
[1] G.-J. De Vreede, R.M. Davison & R.O. Brigss, How a silver bullet may lose its shine. Comm. ACM 46:8, August 2003, 96-101. http://doi.acm.org/10.1145/859670.859676
Of course, I can see the benefits of - for example - computer counting votes or other manual things that would take plenty of time if conducted manually. But if I imagine a system that would do that or other similar simple tasks, it hardly seems like a system that would need a facilitator to run it. Of course, this is all just thoughts based on how I imagine GSS's work. I don't have any personal experience with GSS's, and I don't really know what kind of tools they provide or how hard they are to use.
Other thing that raised a thought was study by De Vreede et al.[1] where they studied causes for GSS meeting failures. Most of the failures had something to do with problems with goals and far less common were the ones with problems with technology. So most of the problems seem originate from the participants, not the tools. Just like in Jami's post here, it seems that it all comes to the people using the tools.
References:
[1] G.-J. De Vreede, R.M. Davison & R.O. Brigss, How a silver bullet may lose its shine. Comm. ACM 46:8, August 2003, 96-101. http://doi.acm.org/10.1145/859670.859676
Thursday, February 10, 2011
Camping at meeting room, experimenting collaborative project work
This week we decided to try something different at office. We had to design company department's web site. The project was not considered extremely critical, and it had long history of postponing and delays. We decided that it was time to get it done.
We decided to arrange something called "meeting room camp". The goal was to finish the project by informally working together. We agreed to gather at meeting room after official work hours and bring our laptops. Some snacks and refreshments were also reserved. Everyone was given the task to design or write the content for some of the pages. The camp started with everyone staring focused on their screen and starting to type.
After couple hours we decided to check the outcome. Everyone had done really nice job, but we found out that the result was far being optimal and lacking coherency.
How this happened! We were not short of collaborative tools or technologies. We had cloud based design tools and shared databases. Remember that we were also sitting at the same room, so we could use all f2f back channels and body language with all their full bandwidth.
It turned out that everyone were so focused with their indvidual tasks that the communication was at least partially forgotten.
After the camp we had little mixed feeling. We did very good progress but the goal was not reached. I'm thinking we hit something fundamental. The collaboration is not always optimal even with optimal tools and environments at hand. My opinion is actually more in people and the habits of doing work than technology.
However, the experience was interesting and now we have the knowledge how to improve the second camp.
We decided to arrange something called "meeting room camp". The goal was to finish the project by informally working together. We agreed to gather at meeting room after official work hours and bring our laptops. Some snacks and refreshments were also reserved. Everyone was given the task to design or write the content for some of the pages. The camp started with everyone staring focused on their screen and starting to type.
After couple hours we decided to check the outcome. Everyone had done really nice job, but we found out that the result was far being optimal and lacking coherency.
How this happened! We were not short of collaborative tools or technologies. We had cloud based design tools and shared databases. Remember that we were also sitting at the same room, so we could use all f2f back channels and body language with all their full bandwidth.
It turned out that everyone were so focused with their indvidual tasks that the communication was at least partially forgotten.
After the camp we had little mixed feeling. We did very good progress but the goal was not reached. I'm thinking we hit something fundamental. The collaboration is not always optimal even with optimal tools and environments at hand. My opinion is actually more in people and the habits of doing work than technology.
However, the experience was interesting and now we have the knowledge how to improve the second camp.
Subscribe to:
Posts (Atom)