[https://github.com/ucl-comp0035/coursework-1-mslee2129]
Assume that requirements will be installed from requirements.txt.
If you have used any libraries that require set-up beyond pip install ... then use this section to explain any set-up
instructions to be followed to run your coursework.
If the marker cannot execute your coursework they can't grade it!
CRISP-DM
Criteria:
- Level of volatility in the requirements
- Availability of documentation/support
- Flexibility of structure and timeframe
- Size of the project team
Justification: For this project, I wanted to focus more on the data science aspect of the project, and CRISP-DM is known as one of the most popular data science methodologies. This would mean that it would have plenty of documentation and support available to assist me throughout the project. Out of the many data science methodologies, I chose CRISP-DM since it is a methodology that allows for the most flexibility, as it is a methodology that combines features of the waterfall and agile methods. It also allows for a number of team members to work on various aspects of the project at the same time, making it an efficient way to organise and distribute tasks. Also, given my experience with data analysis, there would be aspects of this methodology that would be familiar to me, which is an added benefit. The shortcomings of using CRISP-DM in this project, however, could be that it lacks some detail that other methodologies do not, and this may lead to some oversight or lack of clarity when it comes to the finer details of the project. The fact that it is a data science methodology may also be a weakness, since it may overlook the software engineering aspects such as deployment of their models.
Doing business: Different countries offer different environments for a business to thrive, and it is often difficult to identify what countries offer the most potential growth of a business. If a company chooses poorly when moving to another country, it may result in severe financial losses. It is essential for a business to choose to expand to a country that would maximise its revenue. This can be especially relevant to newly expanding companies that want minimise the risk of failing in their new enterprise. Given a set of 205 indicators and how a country scores in those indicators across the years, recommendations will be made on the next steps to take for a business.
See Persona.pdf
- Which countries rank the highest in the 'ease of doing business' criteria?
- Which countries have the most business-friendly tax regulations?
Data Preparation In the data_preparation.py file, I have written code that would clean the data from the DBData.csv file. This was done mostly for two purposes: to filter out certain values in the 'Country Name' column that were not included in the DBCountry.csv file (values such as Beijing, Delhi, etc.), and to filter out the indicators that won't be used in the project. I have also decided to leave only the recent data points (2016 onwards) since the availability of some values prior is inconsistent. Once the cleaning was finished, I exported the dataframe as a csv file called DBData_clean.csv
DBData.csv DBCountry.csv
For the data exploration phase, I have decided to visualise the data using bar charts, as the most useful data points in the dataset are the rankings (of ease of business, paying taxes). These two values are also what would be most helpful in answering the questions identified above. Graphs of the two values is saved as 'rank_ease_of_doing_business.png' and 'rank_paying_taxes.png', respectively. For more detail, the create_line function can be used to plot line graphs for individual countries across the years. The function requires a country code, which is available in the DBCountry.csv file. An example is given as AFG.png. Data Exploration
What I did in the last week: In the previous week I've set up my GitHub account and installed PyCharm to set up my development environment. After the PBL session, I have also taken a look at the different development methodologies and have shortlisted a few methods I would like to implement into my project. I have also taken a look at my dataset I would use in my project to familiarise myself with the process.
What I plan to do in the next week: In the next week I would like to specify which methodology I would use in my project, and focus on the business aspect of the project, including target audience and defining the problem.
Issues blocking my progress (state ‘None’ if there are no issues): There are no problems I have encountered yet.
What I did in the last week: In the past week I have developed my problem statement that is specific to the dataset I will be using by following the template provided in the PBL session. I have also started on creating a persona, which will act as my target audience moving forward. I have also started committing changes to my GitHub repository.
What I plan to do in the next week: In the next week I will iron out my persona, since I will need to add more details such as his values and goals, then upload these files to my GitHub repository. I will then read up on the content regarding data ethics for the next week.
Issues blocking my progress (state ‘None’ if there are no issues): I have not encountered any problems yet.
What I did in the last week: In the past week I was able to finish up writing my persona, as well as get started on observing my dataset. I have followed the PBL materials to familiarise myself with the Pandas library to begin analysing my dataset and test out the different functions.
What I plan to do in the next week: In the next week, I will continue to work on my dataset using the Pandas library to see if there are any irregularities that may cause an error and start cleaning the dataset.
Issues blocking my progress (state ‘None’ if there are no issues): There are currently no issues I have encountered.
What I did in the last week: In the past week I was able to follow along with the data exploration process, which involved cleaning more of the data and starting to visualise some of the data that are relevant to the project. I have worked more with dataframes and the pandas library, as well as trying out some other libraries such as matplotlib.
What I plan to do in the next week: In the next week I should be able to finish up the project, since the deadline is on the 10th of November. I will achieve this by adding more code to my files and committing them on GitHub.
Issues blocking my progress (state ‘None’ if there are no issues): I have not committed some of my work to GitHub as I prefer to work locally and make sure there are no error before, which may have hindered my progress a little bit. However, this was not too much of a problem.
Use any referencing style that you are used to using in your course.
Most students will use the same repository for coursework 2. You may use this file to present the results of that coursework if you wish. Alternatively you can use video or audio to provide the explanations instead of writing them.
The term 'requirements' is used in the broader sense, user stories and/or use cases may be used.
For this coursework, I have decided to use user stories to specify the requirements. This includes the 'As a, I want, so that' format, so that I could capture the target audience and their needs. There are a total of 4 different user groups: CEO, Data Scientist, Marketing Staff, and General Public. The requirements have been given a priority using the MoSCoW method, which classifies the requirements in 4 tiers of 'Must have', 'Should have', 'Could have', and 'Won't have for now'. The requirements and their priorities have been sorted in a table to make it easier to comprehend.
Link to the full list of documented and prioritised requirements.
See Requirements
See Wireframe
For the initial design of the app, I decided to create a wireframe to help visualise the requirements set in my Requirements Elicitation. This wireframe was designed using Figma, an online designing software. Some requirements (i.e. those with lower priority) have not all been included in the wireframe. I have decided to design the app for desktop use, since my target audience would mostly be office workers who use a computer more than their phones for work. The screen size for the wireframe was therefore set as 1080x768.
See Routes, Views, Controller Functions
The architecture has been designed using UML diagrams via LucidChart. UML diagrams are an efficient way to visualise classes, their attributes and methods, and their relations to other classes. The class diagram includes notations of data types, inheritance, and multiplicity to further specify the roles and relations of the different classes created. The routes, views, and controller functions of the architecture is specified in a separate document, with each route having an associated view (using the wireframes created) and controller functions.
See Relational Database Design
The relational database has been designed using an entity relationship diagram (ERD) via LucidChart. The ERD holds information on the entity, keys, fields, and data types, which are all key components of a database. The relationship between entities is visualised using the Crows Foot style.
pytest
The tests should be in a separate and appropriately named file/directory.
Provide evidence that the tests have been run and the results of the tests (e.g. screenshot).
See User Test
Consider using GitHub Actions (or other) to establish a continuous integration pipeline. If you do so then please provide a link to the .yml and a screenshot of the results of a workflow run.