Time Series Analysis & Forecasting Using TM1 Empowered by R (Part 1)

Time Series Analysis & Forecasting Using TM1 Empowered by R (Part 1)

Cognos TM1-R Architecture:

Contents:

1.Overview of IBM Cognos TM1(Part 1)
2.Exporting historical data from Cognos TM1 to R (Part 1)
3.Overview of R (Part 2)
4.Importing historical data into R (Part 2)
5.Performing Time series analysis and projecting forecast in R (Part 2)
6.Importing forecast data to Cognos TM1 (Part 2)

1.Overview of IBM Cognos TM1:

IBM Cognos TM1:

  • Cognos TM1 is an enterprise planning solution for planning, budgeting and Forecasting as well as analytical and reporting applications.
  • It is in-memory tool, it reads/writes data from systems RAM.
  • It stores the data and gets the data from System RAM.
  • It supports cube structure and performs slicing and dicing operations.
  • It address your business challenges

                         i)Structuring

                        ii)Automating process

TM1 Architect:

  • TM1 Architect is one of the standard TM1 client components that can connect to the TM1 server.
  • A complete TM1 model is build in TM1 Architect using Turbo Integrator processes.
  • It let’s you create, analyze and manage business data using the Cube Viewer.
  • Administrative capabilities are only available through TM1 Architect Server Explorer.

Model Building in IBM Cognos TM1:

  • Dimensions
  • Cubes
  • Rules
  • TI Processes

Dimensions:

Dimensions in TM1 are basic building blocks of cube. You create a cube with dimensions, and dimensions identify how to organize the information or data that you want to track and report on. Each element in each dimension identifies the location (or “x-y coordinate”) of a cell in a cube.

  • Different ways to create Dimensions:
  1. Manually
  2. Using Turbo Integrator Process
  3. Using dimension work sheets

Cubes:

  • Cubes of tm1 are like table to relational database. Cubes are collection of dimensions.
  • Almost all data stored in tm1 is stored in and accessed from cubes.

There are two ways to create cubes:

  1. Empty cube: You can create an empty cube by selecting (at least) two dimensions from the list of existing dimensions in the creating cube window to create a new cube with no data.
  2. External data sources: You can create a cube and load it with data by using Turbo Integrator each and every TM1 cube must have at least two dimensions and a maximum of 256 dimensions.(Which is an ETL tool for TM1) to identify and map dimensions and data from an external data source to a new or existing cube.
  3. Each and every TM1 cube must have at least two dimensions and a maximum of 256 dimensions

 

TM1 TurboIntegrator Process:

  • A TurboIntegrator process contains a script of TurboIntegrator functions and commands to programmatically import data as well as create and modify TM1 objects, such as cubes and dimensions.
  • An IBM Cognos TM1 administrator creates the TurboIntegrator process and saves the process on a IBM Cognos TM1 server.
  • The administrator also assigns security privileges to the TurboIntegrator process.
  • One must have read access privileges to access the TurboIntegrator process from the source tree.

TM1 Web:

  • IBM cognos TM1 Web is a client interface where the TM1 cube data is brought from TM1 worksheets through Applications.
  • The screenshot below refers to forecasting cube data. The Hyperlink shown in the screenshot generates forecasting values for next 12-periods.

 

2.Exporting historical data from IBM Cognos TM1 to R

The Excel Input data exported from IBM Cognos TM1 to R:

  1. This is the format of data required for time series needed to export to R
  2. The view need to be created in Cognos TM1 and exported to excel. You may directly read from excel or you can convert into csv file.
  3. We can add an action button in excel to export data to external folder. We have used this method in our sample below.

Conclusion :

In this  blog we have discussed about  the procedure  of time series data and how to export data from TM1  to R.  In the next series of  blog  we will discuss  about R language  and time series forecasting in R .

Leave a Reply

Your email address will not be published.