Part-time Data Science and Full-stack students’ Capstone projects batch #2 and #3
by Guest
The end of a course is always a thrilling time, especially when it comes to showcasing the final projects of our part-time data science and full-stack students. These students have put in a lot of effort to apply the skills they've acquired to solve real-world problems, and their projects are a true testament to their commitment and inventiveness. From building dynamic web applications to developing machine learning models, our students have tackled diverse challenges, and we're excited to share their achievements with you. Join us as we explore the fascinating world of data science and full-stack development and get inspired by the impressive work of our graduates.
Established in 1780, the Neue Zürcher Zeitung remains one of the most-read German-language newspapers in Switzerland, with over 1.5 million registered users and over 200k subscribers. With readership not slowing down as print media is increasingly adapting to the digital domain, there remains a unique opportunity to improve content delivery, giving people what they want to read, when they want to read it.
By analyzing the behavior of readers and performing user segmentation to identify key behavioral groups, together with a better understanding of the news cycles and informational trends of published articles, the interactions between users and articles hold a wealth of insight that will drive the engine of a smart article recommendation system driven by the data.
Reader segments are not simply delineated by the subject matter of the articles they enjoy, but also by primary device usage, the time of day (or day of the week), the length of the articles, and importantly, receptiveness to push notifications via mobile apps or social media. By comparing the similarity of readers, articles can be recommended based on the readership of other users within the same segment in what is known as collaborative filtering.
To further bolster this, the content of the articles themselves can also be analyzed, exploiting the content-based filtering approach to recommendation systems to create a hybridized approach. Using the BERTopic algorithm, articles can be broken down into “topics” - important keywords that highlight what sets them apart from other articles - and then clustered by similarity to find related articles regardless of the original department that published them or the title of the article.
Together with metrics that are calculated to determine article performance, a simple classification model was trained to predict what sort of articles are more likely to be “successful” (in terms of engagement), especially when delivered to the appropriate readers.
This combinatorial approach of user readership patterns and behaviors, article content and metadata, and predicted article success based on historical data allows for the synthesis of a sophisticated, holistic recommendation system suited for the continually digitizing media landscape of today.
As a pet owner, it can be challenging to find the right dog groomer who understands your furry friend's unique needs. Many dog groomers rely on traditional forms of communication, such as phone calls or word-of-mouth referrals, which can be time-consuming and inefficient for both the groomer and pet owner.
Groomify is a platform that aims to solve this problem by connecting dog groomers with pet owners through an easy-to-use digital platform. The platform offers a location-based search option that displays a list of dog groomers in your area. Pet owners can browse through the groomer profiles, read reviews from other customers, and choose a groomer that suits their needs.
Current features:
For the Dog groomer's access side, the team not only built a simplified online platform, but also already included all the essential features needed. This includes the creation of a range of services unique to the specific dog groomer, an appointment management system as well as scheduling options.
For pet owners, Groomify offers a convenient way to book appointments for their furry friends. They can create a profile for their dog and choose from a range of services, such as baths, haircuts, or nail trimming. The platform allows pet owners to schedule appointments that fit their busy lives.
Future features:
Our Groomify team already has plans to improve the platform. These mainly focus on the dog groomers as they want to improve the creation and customization of services and also include an accounting system so the dog groomers have a complete solution to all their needs. For the dog owners, they want to improve the usability of their platform by focusing on creating an easy-to-use app so they can book their appointments on the go as well.
As an Amazon shopper, it can be challenging to keep track of the prices of products that you are interested in buying. Amazon's prices can change frequently, and it can be tough to know when a product has dropped to your desired price range. That's where Amazoup comes in.
Current features:
- Creation of thematic collections to keep track of products.
- Automatic, daily scraping of price data off amazon.
- Email notification when a product price meets the target price.
The Amazoup platform tracks the prices of products on Amazon every day for you. To use Amazoup, you create collections with a target price. A collection is a group of products sorted by topics, such as screens or snowboards. To add a product to a collection you simply enter the link. Amazoup scrapes Amazon daily to get the price of each entered product, saving, and displaying the development in a graph. If the price of one of the products hits the target price, you will be notified by email.
Amazoup is an excellent tool for Amazon shoppers who want to save money and keep track of price fluctuations. By monitoring the prices of products, you want to buy, Amazoup helps you avoid overpaying and alerts you when a product falls into your desired price range.
Future Features:
The Amazoup team has big plans to improve the platform in the future. They want to give users more control over scraping intervals, such as once per hour for sales days like Black Friday, to help users take advantage of limited-time deals.
Additionally, they want to add more graphs to provide users with more options to compare price development. With additional data visualization options, users can make more informed purchasing decisions.
Amazoup is planning an auto-order feature that will order the product automatically if it falls below your target price. With this feature, users can save even more time and effort when shopping on Amazon.
Have you ever wanted to exchange items with someone but didn't know where to start? With Gimme, users can exchange items with each other, creating a bartering community. Instead of selling items, users can create offers of items they want to barter, including information such as the item description, its condition, and location. In addition, they specify what items they would like to trade for the item they are offering.
Current features:
The platform is easy to use, allowing users to browse all the offers and search through them by topic. The platform's messaging system allows users to contact the offeror directly if they find an offer, they are interested in. Gimme also offers users the option to create requests, giving users a way to let people know what kind of items they are interested in receiving.
Future features:
As part of its commitment to continuous improvement, the Gimme team has some exciting plans for the future. One feature currently in development is location-based search, allowing users to find offers and requests that are closer to their area.
In conclusion, our part-time data science and full-stack students are some of the most dedicated and passionate learners we've had the pleasure of teaching. We wish them great success in their future endeavors.
If you're interested in delving into the world of data science or honing your skills as a full-stack developer, we've got you covered. If you're interested in our part-time data science course, you can click here: https://learning.constructor.org/data-science/part-time. Alternatively, if you're curious about our part-time full-stack bootcamp and want to learn how to code and build web applications, you can check it out here: https://learning.constructor.org/full-stack/part-time. We're excited to help you achieve your goals and develop the skills you need to succeed in today's ever-changing digital landscape.
NZZ: User and content-aware article recommendations
Bootcamp: Data Science
Students: Bea Schröttner, Claudia Annoni, Dominik Batz, Renato FillinichEstablished in 1780, the Neue Zürcher Zeitung remains one of the most-read German-language newspapers in Switzerland, with over 1.5 million registered users and over 200k subscribers. With readership not slowing down as print media is increasingly adapting to the digital domain, there remains a unique opportunity to improve content delivery, giving people what they want to read, when they want to read it.
By analyzing the behavior of readers and performing user segmentation to identify key behavioral groups, together with a better understanding of the news cycles and informational trends of published articles, the interactions between users and articles hold a wealth of insight that will drive the engine of a smart article recommendation system driven by the data.
Reader segments are not simply delineated by the subject matter of the articles they enjoy, but also by primary device usage, the time of day (or day of the week), the length of the articles, and importantly, receptiveness to push notifications via mobile apps or social media. By comparing the similarity of readers, articles can be recommended based on the readership of other users within the same segment in what is known as collaborative filtering.
To further bolster this, the content of the articles themselves can also be analyzed, exploiting the content-based filtering approach to recommendation systems to create a hybridized approach. Using the BERTopic algorithm, articles can be broken down into “topics” - important keywords that highlight what sets them apart from other articles - and then clustered by similarity to find related articles regardless of the original department that published them or the title of the article.
Together with metrics that are calculated to determine article performance, a simple classification model was trained to predict what sort of articles are more likely to be “successful” (in terms of engagement), especially when delivered to the appropriate readers.
This combinatorial approach of user readership patterns and behaviors, article content and metadata, and predicted article success based on historical data allows for the synthesis of a sophisticated, holistic recommendation system suited for the continually digitizing media landscape of today.
Groomify: Connecting pet owners with local dog groomers
Bootcamp: Full-stack Web Development
Students: Alessio Tortora, Martin Pöhl, Viktoriia Palii, Angelos ZaimisAs a pet owner, it can be challenging to find the right dog groomer who understands your furry friend's unique needs. Many dog groomers rely on traditional forms of communication, such as phone calls or word-of-mouth referrals, which can be time-consuming and inefficient for both the groomer and pet owner.
Groomify is a platform that aims to solve this problem by connecting dog groomers with pet owners through an easy-to-use digital platform. The platform offers a location-based search option that displays a list of dog groomers in your area. Pet owners can browse through the groomer profiles, read reviews from other customers, and choose a groomer that suits their needs.
Current features:
- Location-based search
- Appointment booking and management system
- Creation of bookable services specific to a single dog groomer
For the Dog groomer's access side, the team not only built a simplified online platform, but also already included all the essential features needed. This includes the creation of a range of services unique to the specific dog groomer, an appointment management system as well as scheduling options.
For pet owners, Groomify offers a convenient way to book appointments for their furry friends. They can create a profile for their dog and choose from a range of services, such as baths, haircuts, or nail trimming. The platform allows pet owners to schedule appointments that fit their busy lives.
Future features:
Our Groomify team already has plans to improve the platform. These mainly focus on the dog groomers as they want to improve the creation and customization of services and also include an accounting system so the dog groomers have a complete solution to all their needs. For the dog owners, they want to improve the usability of their platform by focusing on creating an easy-to-use app so they can book their appointments on the go as well.
Amazoup: a tool to tracking Amazon prices
Bootcamp: Full-stack Web Development
Students: Victor Herrero, Nattira Faerber, Michael Zolliker, Lucio BonforteAs an Amazon shopper, it can be challenging to keep track of the prices of products that you are interested in buying. Amazon's prices can change frequently, and it can be tough to know when a product has dropped to your desired price range. That's where Amazoup comes in.
Current features:
- Creation of thematic collections to keep track of products.
- Automatic, daily scraping of price data off amazon.
- Email notification when a product price meets the target price.
The Amazoup platform tracks the prices of products on Amazon every day for you. To use Amazoup, you create collections with a target price. A collection is a group of products sorted by topics, such as screens or snowboards. To add a product to a collection you simply enter the link. Amazoup scrapes Amazon daily to get the price of each entered product, saving, and displaying the development in a graph. If the price of one of the products hits the target price, you will be notified by email.
Amazoup is an excellent tool for Amazon shoppers who want to save money and keep track of price fluctuations. By monitoring the prices of products, you want to buy, Amazoup helps you avoid overpaying and alerts you when a product falls into your desired price range.
Future Features:
The Amazoup team has big plans to improve the platform in the future. They want to give users more control over scraping intervals, such as once per hour for sales days like Black Friday, to help users take advantage of limited-time deals.
Additionally, they want to add more graphs to provide users with more options to compare price development. With additional data visualization options, users can make more informed purchasing decisions.
Amazoup is planning an auto-order feature that will order the product automatically if it falls below your target price. With this feature, users can save even more time and effort when shopping on Amazon.
Gimme: A Platform for Bartering Items
Bootcamp: Full-stack Web Development
Students: Zahoor Novman, Aleksander RistovskiHave you ever wanted to exchange items with someone but didn't know where to start? With Gimme, users can exchange items with each other, creating a bartering community. Instead of selling items, users can create offers of items they want to barter, including information such as the item description, its condition, and location. In addition, they specify what items they would like to trade for the item they are offering.
Current features:
- Creating offers and requests
- Browsing through the offers and requests with a category- and text-based search
- Easy-to-use messaging system to notify a user that there is an interest in their offer
The platform is easy to use, allowing users to browse all the offers and search through them by topic. The platform's messaging system allows users to contact the offeror directly if they find an offer, they are interested in. Gimme also offers users the option to create requests, giving users a way to let people know what kind of items they are interested in receiving.
Future features:
As part of its commitment to continuous improvement, the Gimme team has some exciting plans for the future. One feature currently in development is location-based search, allowing users to find offers and requests that are closer to their area.
- Another plan in the works is to create a mobile app version of the platform, making it even more convenient for users to barter items on the go. Additionally, the team aims to implement multiple languages, making Gimme accessible to a wider audience.
- Gimme makes bartering easy, accessible, and fun.
In conclusion, our part-time data science and full-stack students are some of the most dedicated and passionate learners we've had the pleasure of teaching. We wish them great success in their future endeavors.
If you're interested in delving into the world of data science or honing your skills as a full-stack developer, we've got you covered. If you're interested in our part-time data science course, you can click here: https://learning.constructor.org/data-science/part-time. Alternatively, if you're curious about our part-time full-stack bootcamp and want to learn how to code and build web applications, you can check it out here: https://learning.constructor.org/full-stack/part-time. We're excited to help you achieve your goals and develop the skills you need to succeed in today's ever-changing digital landscape.
