Predictive Analytics

[qv_testing]

Hi,

1. What is Predictive Analytics?

2. How is it done? can we do it through Qlikview?

3. Whar are areas where Predictive Analytics can be done?

please any one can explain....

Thanks & Best regards,

Raju

[er_mohit]

On-Demand Webinar: Business Discovery & Predictive Analytics using QlikView

Predictive analysis is just as what if analysis

for Sales Company HR module Quality modules you can use it to getting the predictive analysis

[qv_testing]

Hi Mohit,

thanks for quick reply..

What kind of data would be needed for Predictive Analytics.

[#4]

Hi,

This is what Wikipedia says: Predictive analytics - Wikipedia, the free encyclopedia

Here is some info in the forum: http://community.qlik.com/docs/DOC-3981

And here is some video material: Text Analytics And Predictive Analytics In QlikView

Regards,

Gerrit

[er_mohit]

See the below link

http://community.qlik.com/docs/DOC-3981

your quantifiable fields like

if in purchase department the higher level employee used this type of analysis because they only want to see the actual sales and quantities per product regards that comparision with last year suppose some products are not being good as compare to last year the profit seems like to less what they do they find the problem where it is then they find the peek point and do what if analysis means they change the price  as to get profit from previous one it helps to  make better decision making system thts why its more useful to get calculation and works fastly as compare to other BI tool

[qv_testing]

Thank U Mohit... It's very helpful...

Already i seen above link, but when i download demo file(Zip file) it show's error...

[christian77]

Hi.

Predictive analysis is the ability to see what will happen in the future, based on the data we have from the past.

One of the simplest is the regression analysis. You can see your income from the 4 past years is 100, 101, 102, 103,... can you tell what you'll have next year?

In a more complex situation, we do that establishing mathematical or theoretical models that explain the data we have from the past. Models are improved as we have more data and we analyze the differences between our model data and the real data.

In ordinary QlikView I think there is only one way to do that. Linear Regression Method. That's why there is another tool mentioned in this post to do predictive analysis.

And right after that, it comes the Statistical Inference.

Why I should know the future if I can't change it?

Analyzing the past allows us to know what things we have to change now in order to get a better future.

Example: Let's put more money in education since we know the good results from the past.

We can make an inference in our future thanks to mathematical model.

[qv_testing]

Hi Christian,

Thanks for sharing.....

[christian77]

Hi.

I don't think predictive and what if are the same.

What if analysis is simpler and it approaches a little more the mathematical or statistical inference rather than prediction. Anyway it's so risky to make inference based on a what if model and I wouldn't do it.

Here are a few differences:

In what if, you change some input variables or fields or some other  imputes  to see what would happen in that case.

In predictive analysis you CAN'T change anything. Can you change your income from the past years?

In what if, the developer has made a model based on whatever he wants. For example, I sell a product for an average price of some €. Now I can change the quantity from 10 to 100 or whatever I want, I just multiply the new quantity or even the price and voila, I see what I want to see.

In predictive analysis more complex mathematical methods are used, as linear regression. It is not just a multiplication. Of course you can build your what if as complicated as you want but is not the same. One thing is for sure: you CAN'T change the quantity or the price of a previous sale or the number of people who was born last month.

When creating a what if analysis, your goal is something like throwing dices to have an approach to reality.

Formal Predictive Analysis improves the precision of the model by years. It's not something punctual, or short term, it's something in a long term, that is comparing model data with real data always.

Salutes.

[qv_testing]

Hi Christian,

i am confusing....

if you have any document file please send me......