IS428 AY2019-20T2 Assign TIAN. MINGZE

Data preparation

We are given 2 datasets, Raw data 2018-03-07 and SMU Library Survey Comments. Unfortunately, these two given datasets are not well structured and cannot be directly used. So, I will mainly focus on data cleaning and data preparation in this section.

Raw data 2018-03-07 SMU LCS data file

There are two worksheets in this raw data. The first worksheet record all response data from all participants. The second worksheet contains all the column name (legend) for the first worksheet.

Issue 1: When I take a closer look into the data, I realized that all the assessments are in measures format instead of dimensions.

Solution: Use tableau to pivot the selected survey assessment values.
      Detail steps 1. Select column from 'I01' to the end. 2. Under the drop down menu select 'pivot'. 3. Rename the Pivot Field Name to Assessments, Pivot Filed Value to score.

Issue 2: The column name (legend) and data rows are in two separate worksheets. Without knowing about each column's name, it is impossible for us to conduct the visualization.

Solution: Use tableau built-in function to join these two dataset together.
      Detail steps 1. Import the raw data into tableau. 2. Drag the worksheets to the connection field and choose left join (SMU on left side, Legend on right side). 3. Under left join, select Assessments = Code.

Interactive Visualization

The interactive visualization can be accessed here:

asfdsdgvdfgfdxhxfgjdhtrxfdhgsdasFSgfhjnvbcxsdaewartfhgbvcxdaerthgfbxvcsdaertehgbxvdfghf

asfdghjghdgsdfcxbghjutrgfdvcxbnghjytrgdf

Interactive Technique	Rationale	Brief Implementation Steps
Filter participant postion with a single value selection list	Provide flexibility for analysts, the analyst can choose the participant postion that they are interested to analyse. The checkboxes is not used in this case because it requires the analyst manually uncheck the unwanted value. And dropdown manu is not intuitive enough in this case. Thus, a single selection value list is perferred.	Use calculation field to convert the position from digit value to plain text (ie. undergrudate, staff ...) Put both calculation field and the 'position' dimension into filter to filter out those useless value (ie. others, exchange...)
B pic	asd

Reference

https://wiki.smu.edu.sg/1617t1IS428g1/IS428_2016-17_Term1_Assign3_Gwendoline_Tan_Wan_Xin