Objective:
Nowadays, most of the property management companies will install cameras around the properties or parking for security purpose. However, this also incurred a cost for hiring employees to monitor the environment; and the cost would keep increasing as the managing portfolio expands. Our client approaches us with such concern, and without a doubt, we are able to provide them with a technical solution.
Modelling and Analytical Process:
By breaking down the video recordings, we have developed a machine learning model to identify the approaching vehicle
After a vehicle is spotted, the model would search its car plate number in the management records
If its number is matched, parking door would automatically open for entrance; or else, the system would signal the staffs to identify the incoming personals.
Summary:
The solution results in much efficient managing process with fewer recruits. On top of that, we can provide enhancement or modification according to future demand.
Task Models:
Our task is to analyze youth trends for customers and determine why our client should target new youth customers for their loyalty program products. Using the data provided, we decided where youth between ages of 14-17 and 18-24 currently spend their money, through what channels they sign up, and where geographically there is an opportunity for further acquisition of new customers.
Methodology:
We assessed all the provided data and ran RFM to cluster the current customers. We defined four segments, understood customer-shopping patterns, and calculated the share of each of the two age groups within each section. We also compared our two age groups against 25+-year-olds provided in the data to understand how the customers evolve in their purchasing behaviour past the youth stages.
Results:
Our results showed youth want to be digitally connected but there is a decline in activations through mobile, indicating an improvement to the platform may be required. We also understand that youth between 14-17 are key customers for our client to focus on, as they will want a product when they turn 18 and other products as they continue in life. Lastly, we recognize the customers in different segments like to shop at different locations/retailers and want more control over their points. Offering them customized choices to pick where to earn more points from will incentivize them to stay loyal to the program. Should mention personalized offers
Objective:
There is a movie review section on our client’s webpage where its users can put ratings & comments on. Our client hopes to generate business insights from this resource and utilize them to enhance its development plan. By analyzing the hidden information in customers’ reviews, they would be able to precisely pinpoint what kind of movie to recommend for a specific customer.
Modelling and Analytical Process:
Data preprocessing to get the foundation for further model development and analysis
LDA models are developed to predict positive and negative reviews from the customers (Sentiment Analysis). Neutral reviews are filtered since they do not provide meaningful information
Model-measurement to ensure it is accurate and interpretable
An Innovative Solution
The Spinning Studio (Toronto) has gone through a tremendous transformation over the past few months, increasing its membership base and extending its reach to the community through social media and a highly competent and charismatic group of instructors. Our motivation behind this case is to explore opportunities by which the company can optimize its revenue and maximize profit by applying some of the methods and frameworks learned in class, mainly: demand diversion, maximum likelihood estimation, bundling and willingness to pay. This type of business has constraint capacity and receives high levels of demand during peak days/hours which may be forcing it to send customers away. During the week, the order is inconsistent and difficult to predict.
The owners have expressed interest in having more control over their pricing strategy and in measuring demand more accurately. They are currently engaged in a trial period with Class Pass, a credit-based fitness membership that gives users access to a variety of different work out classes (yoga, spinning, boxing, dancing, etc.) across the city with real-time availability for easy scheduling. AS part of this project, we’d also like to understand if this sort of partnerships is necessary for the business, given that this Spinning Studio loses some control over prices set for specific days.
Our goal is to come up with the most effective pricing framework and provide the client with a few options that yield more control on capacity management, resource allocation and certainty over revenues.
Objective:
Our client has installed sensors across their shopping mall to track customer movement for research and potential business needs. They have approached us with 3.4 million of records and hoping to get insights in how to increase customer dwell time and, therefore increase revenue. At first glance, this is a complicated task since we have minimal information on customer background as most of the data only contains stores they have visited. However, we developed the solutions below by utilizing our expertise.
Modelling and Analytical Process:
It is essential to create customer portfolios by identifying distinguishable traits
Machine learning models are developed to flag customers who are potentially leaving