Case Scenario

Solent Intelligence (SOLiGence) is a leading financial multinational organisation that deals with
stock and shares, saving and investments. The organisation operates an online investment
platform that accommodates millions of subscribers (clients) with over 150 billion pounds worth
of investments. SOLiGence is a secure organisation that trades on multiple stock exchange
platforms such as FTSE 100, equities, cryptocurrencies and other commodities. Due to the large
customer base and competition from newly sprung up fintech organisations, the company intend
to implement an Intelligent Coin Trading (IST) platform with an emphasis on crypto coin
predictions. The IST platform will perform intelligent trading by allowing purchases of
CRYPTOCURRENCIES at lower prices and sells at higher prices to realise a substantive profit.
The platform will perhaps anticipate prices of coins on a daily, weekly, monthly and the
maximum of a quarterly basis. (i.e., the system will identify crypto coins with the potential of
buying low and selling high at the specified interval). For instance, a registered user may only ask
to know the coin that will provide a specified amount of profit at a specified interval. For
example, the user may want to know a trade(s), crypto coin (Bitcoins, Filecoin, Ethereum etc)
that will fetch £10 profit today or £300 in profit a month. Therefore, the system should suggest a
particular coin or group of coins that will fulfil the set profit margin.

The correctness of the system’s prediction will be tested with real-world data, and you are
expected to demonstrate the percentage correctness of the system. For example, the system's
prediction for yesterday can be validated with today’s data, and prediction for last week could be
validated with this week data etc.

Your task
You have been employed as a Data Science/AI expert with strong expertise in Artificial
Intelligence (machine learning, precisely) to develop a dummy system that will allow the
company to test and evaluate the sustainability of this intended platform. You have been provided
cryptocurrency data for about 100 different coins, but you are free to utilise data from any
existing stock/cryptocurrency exchange data platform such as Coinbase, Bitpanda, Binance,
Coinbase, Coinmarket and IG amongst other platforms to source for real-time or persistent data
for your model.

The assessment entails the identification/development of machine learning model(s) suitable for
the development of a Graphical User Interface (GUI) software system, i.e., a visualised decision-
making support system for cryptocurrency trade presented with a written individual report that
underpins and evaluates the development process.

The GUI system is expected to suitably help test the company’s intended proposition, but not
limited to the following requirements.

System requirements
 The system should use appropriate data visualisation mechanisms such as graphs and
charts to aid user-system interaction i.e., intuitive data representation.

 The system should be a decision-making support system that will assist users to make
decisions on the intended scenario, for instance, users may specify the interval (day or
week) or profit or coin they intended.

 The system may be a standalone system or online based (this could be Desktop
application, Web-based or Mobile) that graphical displays key performance indexes (KPI)
and trends for each of the key areas.
1
Policy, Governance and Information
June 2022
  You will be expected to apply suitable Machine Learning techniques and algorithms with
justification of their suitability for the system. You may combine multiple machine-
learning models and algorithms to evidence sophistication.

 The system should be interactive and able to handle real-time information, i.e., any
alteration to the data from the interface must reflect immediately. For instance, the change
in price or currency should reflect in the predicted value.

 Your adopted model(s) is expected to address all areas of business operations

determinable from the dataset and case scenario.

 The system should allow the users to plan and forecast investment using a range of
scenarios, i.e., the system permits alteration to answer ‘What-If’ questions (for instance,
what will be the profit if bitcoin is sold for $10000 or $30000. Users should be able to
change the quantity of the coin etc).

Data
The development of an effective predictive model for a system requires a considerable amount of
data that is representative of the intended case problems. Therefore, you are required to utilise the
provided data or any similar cryptocurrency data to develop your system. You will be provided
data of up to 100 different cryptocurrencies for a given interval to help develop your model and
the succeeding system. The data required for learning should contain variables including date or
timestamp, cryptocurrency name, symbol, open, close, high, low amounts. You may want to
utilise a real-time data connection instead of a local data source to demonstrate sophistication.

Questions
After concluding on a suitable underlining machine learning model(s) for your developed system,
your system should consider providing befitting answers to questions and aim to carry out the
following:

 Display a graphical representation of the price over a specified interval for a chosen
cryptocurrency.
 Chose a crypto coin such as Bitcoin or Ethereum and display at least ten positive and
negative correlated cryptocurrencies.
 Display the moving average of a chosen cryptocurrency (you may want to utilise the
timestamp-persistent data or real-time data connecting to a real-world data source.
 Display a predictive time for possible high and low of a chosen cryptocurrency
 Display the best time to purchase and sale of a chosen coin with anticipated amount of
profit and loss.
 Display a prediction of when a group of chosen cryptocurrencies will go up or down
perhaps, the crypto market state.
 Display the confidence level (i.e., the accuracy level of predicted value - profit or loss for
a chosen cryptocurrency).
 Display top stories i.e., RSS feed, about specific or chosen cryptocurrencies.

Software
The choice of suitable machine learning algorithms for the development of software applications
is at the student’s discretion, but as a minimum, the system MUST consist of a data source
(persistent or real-time), which graphical outputted data is sourced. Also, the choice of a front-
end development is at the discretion of the student, but an elegant dashboard format offering a
reliable data visualisation is expected. Student should utilise/build on the IT/programming skills

2
Policy, Governance and Information
June 2022
already developed on the course to develop a concrete front and back system. Students are free to
utilise any comfortable programming language (HTML, Python, Java and JavaScript among
others), particularly for the front-end of the application. It is expected that students utilise a
specific development methodology for designing, implementing, testing and managing the
software (these should include a set of user instructions if not been included in any ‘Help’
menus).

Interface
The system interface should improve user's interaction with the system. The interface must gear
towards the system’s needs/requirements, allowing users to achieve optimal interaction with the
system. Therefore, the system MUST be intuitive, i.e., simple to operate by a new user.

Report
A formal technical report of approximately 2500 words that critically evaluates the developed
model and system (i.e., explaining the process of development) should be enclosed with the
submission. To critically evaluate the developed model/system, emphasis should be made in
tandem with existing theoretical models. Evidence of understanding through critiques of existing
concepts in literature MUST comprehensively reflect in the report. Also, any adopted model(s)
should be well justified, particularly about their suitability for the problem case, the challenges
you undertake and the issues concerning its appropriateness.
You will be expected to discuss problems identification, data dimensions, pre-processing,
models’ suitability and system development; your understanding of machine learning principles
and concepts and its applications for system development and decision-making will be
considered.

Demonstration and Presentation

To demonstrate an understanding of machine learning concepts and principles, and the process of
development, individual exhibition/presentation slot shall be communicated by the tutor as
appropriate. Students should prepare to discuss problem case analysis, challenges, model
evaluation and process of development with an exhibition of the developed system (a simple
PowerPoint presentation may help). The demonstration is equally a critical aspect of the
assessment; thus, it is important to attend this session as it aims to help you and your tutor have a
better understanding of the submitted portfolio. Your tutor shall provide feedback in this session
to allow you to reflect on your work and in developing better systems and conducting effective
research in the future.

Appendices
While the contents of appendices do not count towards the word limit, the section may include
detailed evidence of the model, functionality and the process of development. This could include
images, system architecture and development guides among other things. It may also contain
bibliographies and references of texts. All materials should be properly cited in text and
referenced as appropriate

3
Policy, Governance and Information
June 2022
4
Policy, Governance and Information
June 2022