Daniel Gaisford

See My Work

Hi, I'm Daniel...

I am a 19-year-old visionary founder with ambition to build an influential company that will positively impact people's lives on a large scale. I am fuelled by relentless determination and desire to innovate, working to leave a lasting legacy on the world.

My Work

E-Commerce Screenshot

E-Commerce

From January to April 2024, I built an E-Commerce brand on Ebay. This was modestly successful but I decided to end it due to scalability constrained by ebay's sellers governance .

24HR Gym Screenshot

24HR Gym

In February 2024 I spotted an opportunity to fill a gap in the market and so until May 2024, I created an in-depth business plan for a 24 hour gym in a local town. Unfortunately, I was beaten to market by Pure Gym LTD who rented the very building I had built my plans around.

View the Business Plan
Video Advertising Screenshot

Vide Advertising

As of today, I am working to merge the benefits of digital marketing with the benefits of Out-Of-Home (OOH) marketing. I will do this by monetising food delivery drivers, providing the venture with a network of digital, mobile ad-space with a vast reach. The USP is my computer vision software which I have built that provides advertises with an analysis of demographics.

Read More View Website

More about Vide

Accurate Targeting to Out of Home advertising (OOH)

True Data to Out of Home advertising (OOH)

Vide Feature Image

Technical Achievements

At the start of the year, I had no prior coding or computing experience. Since then, I have successfully developed a full-stack platform for programmatic, dynamic digital out-of-home (DOOH) advertising. The platform integrates seamlessly with both server and client devices, enabling the dynamic display of adverts while capturing detailed metrics on the quantity and demographics of viewers for each ad shown. This data is then presented through a user-friendly interface. Additionally, by leveraging a mobile network of ad spaces, the platform enables unprecedented targeting capabilities in the DOOH advertising industry.

From Zero Coding/Technical Experience

*In 6-
Months*

To Working Prototype Running Fine-Tuned Vision Models

Scooter

Traffic Analysis

From engaging people in the OOH advertising space, I have found that there is clearly a need for improved analytics for advertisers. I have created a system to count the number of impressions advertisements receive. The advertising box has three screens with a camera placed above them. Using computer vision software, we count impressions and gather insightful data on viewers. This system collects data that was previously only available to advertisers through online or digital marketing.

Traffic Visualization

Fine-tuning Computer Vision Model

A problem I encountered was that my software was counting parked vehicles and vehicles traveling in the opposite direction. These vehicles either didn’t have any people in them or had occupants who likely wouldn’t see the advertisement, so they shouldn’t be counted. How do you teach a model to differentiate between a parked car and a moving one? How do you train a model to distinguish between a car passing in the opposite direction and one traveling in the same direction?

To overcome this problem, I fine-tuned a vision model that can identify the lane the camera is currently in. Using this information, we only register vehicles that significantly overlap with our lane. This effectively ensures that only vehicles moving in the same direction and not parked on the side of the road are counted.

Road

Demographic Classification

To allow for demographic targeting, I used classification models to constantly analyse and record demographics. By having a database containing data of which demographics are where at what time, we can predict where and when we can find the said demographic at the highest concentration; this allows advertisers to reach their desired audience. The models are capable of age, gender, ethnicity and emotion classification. By marking key points on individual’s heads, we can even track where they are looking and for what period. Of course, no images are saved to uniquely identify individuals as this would violate GDPR regulations.

Dashboard

Interactive Dashboard

I built a frontend for users to create campaigns, set variables and view results. I built this using HTML, CSS, vanilla JavaScript and React JS. This frontend is all connected to a server and to the motorbike boxes, leading to a functional, practical user experience.

My *Current* Ambitions

Social Media like Targeting in OOH Marketing

The Out-Of-Home advertising market is currently under-served when it comes to data and targeting, this forces many advertisers to prioritise digital marketing over OOH, this is unfortunate as advertisers loose out on the many benefits provided via OOH marketing. I plan to solve this issue using computer vision.

Free Food delivery

With autonomous vehicles and advertisement, I envision delivery for food, medication, and other short-hall logistics becoming free for consumers. With Uber Eats having a valuation of $20bn, the addressable market is large.