IBM Cognos Analytics Integration with IBM SPSS Modeler Subscription:

[CognosBIAnalytics]

Hello All,

I have created a Dashboard using IBM Cognos Analytics on Cloud/Web. Now for Predictive Modelling I need to integrate with IBM SPSS Modeler Subscription
I am trying to use IBM Cognos Analytics as Source from Source Tab: Clicked on Connection: Edit. After entering the Cognos Server URL its producing an Error:
It says: Failed Connected to Cognos Analytics Server.:

Do we need to enter any port number or any other path in Cognos Server URL
Below is the Cognos URL I have given: https://cognosnext.bi.ibmcloud.com/bi/

Let me know your thoughts on same: Do Advise..

Regards
Cognos BI

[MFGF]

Hello All,

I have created a Dashboard using IBM Cognos Analytics on Cloud/Web. Now for Predictive Modelling I need to integrate with IBM SPSS Modeler Subscription
I am trying to use IBM Cognos Analytics as Source from Source Tab: Clicked on Connection: Edit. After entering the Cognos Server URL its producing an Error:
It says: Failed Connected to Cognos Analytics Server.:

Do we need to enter any port number or any other path in Cognos Server URL
Below is the Cognos URL I have given: https://cognosnext.bi.ibmcloud.com/bi/

Let me know your thoughts on same: Do Advise..

Regards
Cognos BI

Hi,

From the URL it looks like you are using the free trial of Cognos Analytics? This isn't something you can connect Modeler directly to in order to pull data out of. You would need your own installed instance of Cognos Analytics if you want to be able to connect to it and import package data into Modeler. Additionally, Modeler can only import from CA packages published in Compatible Query Mode, but CA on the cloud (including the free trial) only supports Dynamic Query Mode, so you can't use Cognos Analytics SAAS as a data source for Modeler directly.

ftp://public.dhe.ibm.com/software/analytics/spss/documentation/modeler/18.0/en/ModelerSPOnodes.pdf

You will probably have more success if you create a list report in CA including the data you want to bring into Modeler, then run and render the report as csv or xls. You can then bring in the csv or xls.

Best regards,

MF.

[CognosBIAnalytics]

yes, I am using IBM Cognos Analytics Cloud Trial Version as of now...
Is there any way we can connect to SPSS Modeller as I need to showcase Predictive Data in Cognos Analytics.

[MFGF]

yes, I am using IBM Cognos Analytics Cloud Trial Version as of now...
Is there any way we can connect to SPSS Modeller as I need to showcase Predictive Data in Cognos Analytics.

As above, you can't connect Modeler to the free trial of CA and use CA as a data source for Modeler. You would need a full version of CA for that. Wouldn't you want the reverse, though? Wouldn't you want to take the predictions from Modeler and use them in CA? Just save your prediction data as csv and import it into CA?

MF.

[BIAnalytics]

yes, thanks.. Yes i did the same and used excel as a source to predictive modelling in SPSS Modeller and again exported to SPSS modeler to integrate to existing Cognos Analytics Dashboards..

[BIAnalytics]

Also i request thoughts on below queries:

1:As of now there is a way we can connect Cognos Analytics as a source with SPSS modeller. Are there any other tools where we can use Cognos Analytics as source and can do predictive modelling. If yes , which are those tools??

2: In case we are forecasting sales/revenue, which is the best algorithm/model which is used across several industries..

[MFGF]

Also i request thoughts on below queries:

1:As of now there is a way we can connect Cognos Analytics as a source with SPSS modeller. Are there any other tools where we can use Cognos Analytics as source and can do predictive modelling. If yes , which are those tools??

Depending on the level of sophistication you require, Watson Analytics might also qualify for this. You can use CA directly as a source for WA, and ask questions in WA such as "what drives satisfaction" or "what is a predictive model for satisfaction" (where 'satisfaction' is an item in your data set loaded from CA). You have no control over the statistical algorithms WA uses for this - it just "does its thing" - it is designed for business users rather than data scientists.

https://www.ibm.com/support/knowledgecenter/en/SS4QC9/com.ibm.solutions.wa_an_overview.2.0.0.doc/wa-an-admin-dataconn-c11_admin.html

2: In case we are forecasting sales/revenue, which is the best algorithm/model which is used across several industries..

I can't give you an answer there, sorry. You'd need input from an experienced Data Science person :)

MF.

[BIAnalytics]

Thanks for your revert....