Week 6

By wiesel |

Monday, August 31, 2009

0 comments

I'm a bad human being. I haven't started the assignment, eek. Been way too busy not doing it. It still doesn't make it okay. I can't decide if i want to do it at home or not. I might download some shiz from dreamspark if i have quota left in my internet usage cap at the end of the month. I really should start.

Le lecture was okay. I apparently am not as good at identifying dimensions as I thought. I made a rookie mistake in the model i was making in my head when POD made us think about the data (pictured). I can't remember what my mistake was but i was a little bit unhappy with myself. I thought i was going to get it right, but i messed it up. I was effected by the overconfidence bias i believe. Which is weird cos when i did the overconfience bias test i got 9 or 10 out of 10 when we were meant to do it with 90% certainty, so i was on the money.

Crappy labs slowed me down, i had to login to another computer once i had already got VS up and running, what a bummer. I am going to be sick of pivot tables by the end of semester aren't i? I see how they are quick and easy, i just don't like them. I found myself using them at work the other day, i wanted to shoot myself then i realised i was saving heaps of time using it rather than writing vlookup statements.

I didn't get up to MDX in the tut, which i was gutted about. That was the bit i wanted to do the most. I got stuck with my cube not loading properly in excel, it was really annoying and it really slowed my progress. I had to refresh the cubes in VS from memory. I tried just redeploying the whole project but that didn't work. Anyways that's over now.

I'm really over my Thursday schedule now though, cos pretty much every week I am busy from 9am-4pm with group meetings and classes and have no time for a break and considering it's week 6 that's not going to change in the next 7 weeks. I think i need to do what happens in Multiplicity.

Week 5

By wiesel |

Tuesday, August 25, 2009

0 comments

Already it is week 5 and weird things are happening in the lecture theater. Sound mishaps and internet dropping out. The weird and wonderful things that happen at 9pm in K-block. POD was complaining that he wasn't coherent. What's new? just kidding.

The lecture was a little like what you said when you subbed in for David Arnott for IT for Management Decision Making with a little bit of stuff about what OLAP was. Seems like POD is on a quest to prove that nothing is new. Although I agree, it's pretty hard to prove with a lot of things and boundaries are blurry. Last night Alex and I were talking about Ford and how they brought cars to the masses. I said something like 'all he did was use Colt's idea for the production line and applied it to cars' and it got me thinking about what invention actually is. Which I shall not get in to now, but it's pretty interesting.

This is probably going to be a short post this week cos i wanna dash off to COSTCO to see what it's like. I went to COSTCO in London and that was an experience.

More handy pivot stuff this week in the tut. it seems like you want to make us experts in Pivot tables. Not sure how much i want to be good at Pivot tables though (see last week's rant). Steve is as excited by everything as always. He and Olga seem to get along very well. The others in the tute are pretty quiet, it's normally infoholic, Olga and I talking most of the time.

Till next week...

Week 4

By wiesel |

Monday, August 17, 2009

0 comments

The pivot week.

It's not that i am an IT snob. Okay, maybe i am. But pivot tables aren't really that intuitive to me. Just cos you can drag and drop doesn't mean that that makes sense. That's leaving alone the facts that you are using a absolutely horrible data structure for decision support. Clearly excel isn't very scalable. You can do everything in the query if you want to (as is in excel) but it's not easier from a user perspective. If you want a usable data structure star schemas are much better. They require extra effort to create rather than just using an excel table but it's a much better solution. You can make a star scheme in access and make queries in that and it should be much easier (evidenced by the FI5095 DW tute in access).

Maybe I have too much trust in the POD but i think i will be following his suggestions on when to do work on the assignment and do it in small bits after we do stuff in the tutes. Which means I need to start doing my assignment, which is really disappointing for me as a lazy person.

Something tells me the diagram on the right is going to be on the exam. I have no idea how i would get that idea. Might have been the incredibly subtle hints, not sure though. I think it's just the vibe. The great thing about this model is that you kind of know it already. Like it's like a combination of everything i've learned in a way. It's a triangle (kinda) and has evolutionary development like Keen's model. It's kinda got SDLC-like boxes in; planning, deliver and use. What goes in those boxes is sort of obvious. The circles are just components (sub-processes) of the boxes. I could argue that strategy and resourcing is typically a little bit going to be controlled in the boardroom and could be outside the boundaries of evolutionary dev. You could add a big box around everything saying organisational culture or something similar, but then it would be ugly. I find that you can put such a big box around almost all diagrams.

You only have 15 readings for week 5. I look forward to skipping them.

Week 3

By wiesel |

Wednesday, August 5, 2009

0 comments

I made this with Crystal Xcelcius:



Pretty pretty isn't it? It was pretty easy to use and to make, easier than with Visual Studio probably. Fairly intuitive for me, it's pretty windows-esc so it's got interfaces that look vaguely familiar and are fairly logically organised. I didn't really have a problem with the tool other than that it's flash, which I don't really like. Linking to flash is just a pain and is horrible for searching and indexing. Bu other than that it was okay. The spreadsheet was pretty damn simply really, and no real forecasting model would be that simplistic if it was any good. But it illustrated Xcelcius (i keep calling it the tool cos of its spelling) very well.

The lecture again seemed to be in the search of O'Donnell's law or should I say O\'Donnell's law as Moodle puts it [Every time i see that i think, shit if they can't clean SQL what hope to assignment submissions have and grade calculations, lol]. I felt like the whole way through the lecture i was being like one of those annoying ppl at magic shows who always asks the Magician how and then has to know how and goes home and is still thinking about it the next day. I was trying to figure out who the companies were in the case studies, the hints were unbearable, 2.5 hour flights, revenue figures, etc, all made me want to go home and look up the historical data for companies at said date. I did attempt to pay attention to what they did wrong or right in their BI system but it was always in the back of my mind that i was thinking about who they could be. It's okay, i think i have given up on trying to figure them all out now, Alex kinda helped a little bit and i resisted the urge to google and some of the hints i've forgotten or forgotten the details of and would have to go back to the lecture recordings to figure it out and that's just not worth the time and effort when I have amazing blogs to write.

The lecture made me think that maybe most BI is disposable, maybe we don't need really fancy systems to do these things, maybe we can just pull all the data together once and get the picture and then move on. Organisational memory is important though, you probably want to retain data, aggregated or otherwise about other times so that you can relate it to the present. Seeing the Development pattern of Case1: A mature system did remind me of Peter Keen's model before you pointed it out, i was feeling all clever after that.

To be honest I don't really rate these surveys you do. I'm not sure if people really tell the truth on surveys even if they are anonymous. They are all perceptions of reality heavily influenced by hindsight bias and groupthink (thanks George Orwell for bringing this term to my consciousness [i read 1984 recently]).

I really don't believe that outsourcing works well for BI, Arnott has me a bit brainwashed here, but it just seems like the kind of activity that you want to be better at your competitors at. If you understand your industry's customers better than your opposition you clearly stand to gain, i'm not sure that if you outsource BI you are going to have an advantage for very long because your competitor will get the outsourced BI system too and then it will be a level playing field. Sure there are different BI outsourcers, but you must be careful about contracts so that the outsourcers can't just sell your system to somebody else. [nb: this is not a well formed argument and i'm not sure i agree with everything i say; end disclaimer]

Week 2

By wiesel |

Sunday, August 2, 2009

0 comments

It's not always interesting listing to somebody else explain the same thing to you that you have heard before. There are some prerequisites for it to stay interesting:
1. You were interested in the first place
2. It's not a carbon copy of what you heard before, ie: it must be of a different format and convey information in a different format, it would also help if there was extra information that you didn't hear before.
3. The delivery and information must be good. If the information and delivery is worse than the first time you are sure to think that it is a waste of time.

Critically for me the lecture this week ticked those three boxes. I had heard a lot of the stuff before but there was stuff spliced in to make it interesting and to keep me entertained. The material was a little bit different to what was the content in IT for Management Decision Making IT4MD).

I think it's funny how Arnott didn't go over POD's paper in IT4MD. Must be cos it's not published and properly peer reviewed, but fudge that. It's still evidence if it's not peer reviewed if the procedures were good.

The idea that I brought away from the lecture was that BI was hard but doable, just like a good woman... SNAP! Stupid immature jokes aside, i think there is a lot of evidence that BI systems can be fantastic when done correctly.

One thing that i would like to criticise is that BI ppl always seem to quote how important BI is in industry surveys of things that they might be considering in the next few years. I think that BI systems are something that most managers don't get around to implementing. I think quoting stats about how likely they are to implement things is a bit of a useless stat. I am planning on clearning my car, i should do it in the next two week and if a survey asked me if i was planning on it I would say yes but I haven't cleaned my car in a while and it's needed a clean for ages and I still haven't cleaned it. It's just not high on my priorities. It's much more important for me to write this blog, do other assignments and watch my football team lose after the siren and then write about it (it's a weird reflective experience; a kin to writing this post).

Secretly... psst... creating that app in visual studio was easy. I think it's cos it's so similar to the music player for mulo things that we created in SDI that it was easy for me. I'd done the ODBC thing before sometime too, I all seemed familiar. Great instructions too, maybe some more screenshots would have been good, there was one part where i was confused and there wasn't one, can't remember. I found that i didn't need to read the instructions, most of it was easy if i skimmed and then guessed.

Infoholic was late, which i communicated to the class brilliantly after he called me once he had acquired my number from the wonderful Alex. Small world. He then proceeded to go through the history of reporting apps in 35mins, which was quite interesting, i was aching to get started on the app a bit, but oh well. I did have time to complete the app though in class, but i don't think my fellow students were as fast as me. I am a bit of a bullet at these kinds of things when i have a bit of experience.

That's probably enough talk for the week, I better steer away from the TV right now. Australia are taking wickets and I have work tomorrow, don't wanna be up all night.