Assignment 3
IBM Cognos Analytics Software Tool for Analyzing Trends in Laptops Dataset Using Decision
Tree Model Development
IBM Cognos Analytics Software Tool for Analyzing Trends in Laptops Dataset Using
Decision Tree Development
What was the Decision Tree Model and why was it good for usage to analyze trends in the
laptops prices? i Generally speaking the Decision Tree displayed a map of the possible outcomes
of a series of choices for the Data Scientist to weigh all the possible actions against one another
based on the best probabilities, costs, and benefits. The Decision Tree allowed the Data Scientist
to focus on finding the trend for the laptops prices by analyzing the dataset. i Also the Data
Scientist examined the correlation among the variables of the laptops prices dataset.
The cleansing of the laptops prices dataset by the Data Scientist has to be performed to
“eliminate mistakes, bad records, data entry errors, and outliers” (Knode, 2019) before the
Comma Separated Value (csv) file can be uploaded into the Cognos Analytics software tool. This
tool served the purpose for “self-service with intuitive web-based interface to upload and analyze
the data,” “supports visual data exploration without writing any code,” “drag and drop data onto
the canvas and made updates to suite analysis,” “the system recommends visualization based on
selected fields and based on currently open visualization,” “you may build interactive dashboards
and stories,” and “show data for an individual visualization” (Knode, 2019).
The Data Scientist began to examine the laptops prices dataset in the Comma Separated
Value (csv) file. Upon close examination, the Data Scientist counted sixty-six records out of a
total 1303 records that have “No OS [Operating System]” data and they were deleted from the
dataset to eliminate a mistake because the Data Scientist should not compare laptops prices with
“No OS” and “OS.” Then there were 1,237 records left. Then the Data Scientist reviewed for bad
records. i The Data Scientist had knowledge and experience with the “Product” column. i There
were 117 duplicate records as part of the “Product” names were deleted because these data
already existed in the other columns. Next item was to look for the typo or mislabel of any
variable. i The Data Scientist spotted the name “Memory” to list the physical hard drive of the
laptops. The correct word was changed to “Storage.”
Then it was time to upload the laptops prices dataset in the Comma Separated Value (csv)
file into the Cognos Analytics software tool. Upon uploading the dataset into the tool, the laptops
prices (Euros) were inserted into the Data slots’ required Target to generate the Tree diagram.
The Data Scientist’s goal was to “find the best set of predictors and optimal way of
combining them so that an optimal model is computed. The insights that are obtained from
decision trees are presented in the form of decision rules where combination of predicators and
corresponding values provide a single prediction for the target value. Decision rules are ranked by
strength so that you can easily find the rules that the most relevant and interesting…the overall
decision tree predictive strength that provides relative improvement the basic model” (IBM).
Given the RAM variable, the Decision Tree algorithm searched the other variables’ data
fields and added them to the model to improve the strength for predicting the target values. The
Data Scientist noticed the Top 5 target values are the following (from number one to number
five): 2667.6, 2259.73, 2017.99, 1827.27, and 1769.7. i The other variables included “Gpu, Cpu,
Screen Resolution and others” to “predict Price Euros with a predictive strength of 74.5%.” The
RAM was the “most important predictor of Price Euros” (IBM). For example, the Data Scientist
examined the RAMs 2GB, 4GB, and 6GB has laptops prices ranges from a low of 174 Euros to a
high of 1799 Euros and an average 575.56 Euros with Std. dev. 271.58, and the total number of
laptops was 406 or 33%. i The next variable for these RAMs 2GB, 4GB, and 6GB to be reviewed
included the CPU and the Screen Resolution.
Then the Data Scientist reviewed the RAMs 8GB and 12GB to have laptops prices ranges
from a low of 389 Euros to a high of 3949.4 Euros and an average 1206.37 Euros with Std. dev.
468.05, and the total number of laptops was 613 or 50%. The next variable for these RAMs 8GB
and 12GB to be looked are the CPU, GPU, CPU, and the Type Name. The Data Scientist’s insight
was that the RAMs 8GB and 12GB provides a strong predictor to buy the good laptops within the
389 Euros to 3949.4 Euros laptops prices ranges from the detail Tree diagram.
Upon further examining the laptops prices dataset, the Data Scientist figured out that the
the RAMs 16GB, 24GB, 32GB, and 64GB have laptops prices ranges from a low of 859.01 Euros
to a high of 6099 Euros with an average 2071.46 with Std. dev. 745.85, and the total number of
laptops was 218 or 18%. The GPU, Screen Resolution, and GPU are reviewed further for these
RAMs 16GB, 24GB, 32GB, and 64GB laptops prices dataset.
Further examination of the laptops prices dataset, the Data Scientist wanted to review the
detail visualization of the Tree sunburst. This sunburst within the decision tree measured the
“strong predictors for a target value or target values, then the colors prevail in that node. The non-
significant values are left out” (IBM). The ranges are from the top target value 2667.6 and down
to lowest target value 438.42 for the Price Euros. i The Data Scientist’s insight was that the RAMs
8GB and 12GB indicate a strong predictor to buy the good laptops prices because total number of
laptops brought was 613 or 50% of all laptops.
Finally, the Data Scientist reviewed the Rules found for the laptops prices dataset. Using
the Predicted value, the Data Scientist was able to find the same Predicted value on the Rules tab
corresponds to the Tree diagram and the Tree sunburst visualizations. For example, the Predicted
value 2667.6 was found to be top target value for all three visualizations: Tree diagram, Tree
sunburst, and Rules.
Conclusion
Data Scientist needed a good analytics software tool like the IBM Cognos Analytics
Software Tool to find out the trends and do the data analysis of the laptops prices dataset. But
before the Data Scientist can upload the dataset into the Cognos Analytics, the dataset must be
examined to “eliminate mistakes, bad records, data entry errors, and outliers” (Knode, 2019). i
Then the Data Scientist reviewed the correlation among the variables of the dataset. The RAMs,
GPU, CPU, Screen Resolution and other variables were used to determine a strong predictor for
the Tree diagram and the Tree Sunburst. i All three data visualizations: Tree diagram, Tree
sunburst, and Rules were found to have the predicted value 2667.6 as the top target value.
