HeaderSIS.jpg

Difference between revisions of "IS480 Team wiki:2018T1 FORD FINALS WIKI"

From IS480
Jump to navigation Jump to search
(update)
(update)
Line 107: Line 107:
  
 
====<div style="background: #f87a54; padding: 15px; font-color: green; line-height: 0.3em;">Task Metrics</div>====
 
====<div style="background: #f87a54; padding: 15px; font-color: green; line-height: 0.3em;">Task Metrics</div>====
[[File:FORD_Midterms_SM.png|center|950px]]
+
[[File:FORD_Final_TM.png|center|950px]]
  
 
<!--Summary of analysis for the metrics collected. You may refer to another page for the details about the metrics and how it is collected.-->
 
<!--Summary of analysis for the metrics collected. You may refer to another page for the details about the metrics and how it is collected.-->

Revision as of 21:15, 14 November 2018

FordLogo.png

Ford home logo.png   HOME

 

Ford about us.png   ABOUT US

 

Ford project overview.png   PROJECT OVERVIEW

 

Ford project management.png   PROJECT MANAGEMENT

 

Ford documentation.png   DOCUMENTATION

 


Project Progress Summary

FORD Final ProgressSummary.png


Final Slides: Final
Deployment Link: [| Chatbot] [| Admin Portal]

To ensure data confidentiality please request the username and password from the team.


Project Highlights

FORD Final Highlights.png

Project Management

Project Status

FORD Final ProjectStatus.png



Project Schedule (Plan Vs Actual)

Planned Schedule

FORD Midterms OldSchedule.png


Actual/Current Schedule



FORD Final Schedule.png


Project Metrics

Bug Metrics

FORD Final BM.png

Task Metrics

FORD Final TM.png


Project Risks

Risk Management


FORD Midterms RM.png


Change Management


FORD change mgmt.png



Technical Complexity

1.Deployment diagram

FORD Midterms TC1.png



2.Chatbot back-end processing logic

FORD Midterms TC2.png



3. Admin Portal back-end processing logic 3.1 We have chosen pandas for our data analytics functions. We have used Pandas DataFrame to load the entire SQL table in the local server memory using only a single, simple SQL select statement. and. This provides the necessary capabilities to transform the data and perform multiple statics functions in a single aggregate function while these data manipulations are done on local memory. After Filtering the data using Pandas, we can store the data into a list and encapsulate the list in JSON format, thus supporting our MVC framework
The advantage of using this approach are as follows and not limited to :
a. Eliminate the need to craft complex SQL statements for future code maintenance after handing over to PSA
b. Eliminate Network Bottlenecks by reducing SQL queries to the SQL server, reserving precious bandwidth for the Chatbot system
c. There would only be one SQL query call to the SQL server per user
d. Any filtering function triggered by the user would be between browser and backend server

FORD Midterms TC3.png


3.2 Hits analysis
a. We have sent the available dates from the backend to frontend by AJAX
b. We query a single select SQL statement to get the necessary data from different tables and store the results into Pandas DataFrame
c. By default, the analysis data is calculated for the current year view, we filter out the columns where the questions are deemed as resolved.
d. We then filtered out the data that are not in the current year
e. We use a single DataFrame function to group the result by the timestamp in the month frequency and to calculate the Question in terms of count and Accuracy in terms of mean.
f. The resolved and unresolved counts are being processed from each individual mean in each row
g. The data are sent to the front end using a list encapsulated in JSON format

3.3 Category analysis (2-tier analysis)
a. Using the same SQL query data previously, we load that data into another Pandas DataFrame.
b. we use the DataFrame function to filter out the questions that are unknown, group the data by subcategory name. We also used the aggregate function to calculate the Question in terms of count and Accuracy in terms of mean.
c. We sort the DataFrame in ascending and convert to list and sent to the front end by encapsulating in JSON format.

3.4 Top K question analysis
a. Using the same SQL query data previously, we load that data into another Pandas DataFrame.
b. We use the DataFrame function to filter out the questions that are unknown and faq questions that don't exist, group the data by FAQ question. We also used the aggregate function to calculate the analysis id in terms of count.
c. We drop the rows of the DataFrame that are of above rank 10 and reset the index.
d. We convert each column in the DataFrame to list and sent it to the front end by encapsulating in JSON format

3.5 Chart.js For Data Visualization
a. We create the rule based on client's expectation to change the statistics and value range of X, Y axis. When the graph filter is changed from the month to the day, the thickness of the bar, as well as the value range, is changed accordingly. Additionally, we add the accuracy rate line to the hit analysis graph, the accuracy rate axis is added to the right side of the graph. This ensures adequate display no matter how data changes, which leads to a better user experience.
b. We update the Chart dynamically. We have used AJAX asynchronous connection to retrieve the available date and pass back the parameters for backend pandas to process the filter logic. Then, we destroy the current charts layer by creating a new layer to display the charts with filtered data that was sent from the backend.


4. Test Automation Framework The test automation framework consists of two parts: Chatbot accuracy rate testing and web application behavior testing. For accuracy rate testing, we checked whether the answers returned from the Chatbot are matching with the expected answers from the test case. For application behaviour testing, we utilized Selenium Webdriver to automate the test cases for our web application. Maven is used to automating the build process, which makes the compilation and packaging more convenient.

Quality of product

Intermediate Deliverables

Stage Specifications Modules
Project Management Minutes Minutes
Metrics Metrics
Risk Management Risk Management
Change Management Change Management
Analysis Use Case Use Case Diagram
Architecture Architecture Diagram
Testing User Testing 1 UT 1 Plan & Results
User Testing 2 UT 2 Plan & Results


Testing


FORD Midterm UAT1.png


FORD Midterm UAT2.png


Click here to access detailed UAT results.

Reflection

Team Reflection

FORD Midterms LearningOutcome.png