How can we utilise AI to validate data from tools like Google Analytics 4?

We work with hoteliers to provide insights into their website and digital marketing performance by aggregating Google Analytics 4 data from hundreds of hotel websites and benchmarking key metrics against industry averages. Of course, we must ensure that the data we compare is relevant and accurate. This is where we face a number of challenges such as outliers, incorrect data, data gaps, and incomparable data, which can skew our analysis if we don’t identify and address them. These issues are compounded by hoteliers implementing Google Analytics 4 in different ways. Using Google’s API, we feed exported GA4 data into a tool called Metabase for analysis. Ideally, we want to review the exported GA4 data (from the API) for inaccuracy/issues so that we only import valid/valuable data into Metabase. Or find a way to interrogate the data once imported into Metabase.

Currently we conduct manual reviews to verify data accuracy which is hugely time consuming and not always the most reliable approach. Our aim is to improve data reliability and accuracy through automation and AI to save time and improve the quality of our benchmark reporting. We believe this could be achieved in a three phased approach. 1) Identify issues/anomalies 2) Identify issues/anomalies and suggest possible fixes 3) Identify issues/anomalies, then suggest even implement the fix (with an ‘ok’ from us) via access to Google Tag Manager/Google Analytics 4.

How can AI enhance our ability to interpret, analyse and strategize hotel performance data reporting?

One of our core services involves providing ongoing website and marketing performance reporting and analysis for hotels, along with insights into broader hospitality and travel industry trends. We integrate data from numerous sources, including GA4, Google Ads, Facebook, Instagram, Ahrefs, Google Search Console, Google My Business, DerbySoft etc. We also consider industry reports from STR, Expedia, and tourism bodies like ABTA. We’d like to explore the use of AI and automation to elevate our reporting capabilities. Ideally, AI would seamlessly ingest data from these sources (via APIs or scraping as required) and deliver high-level, actionable insights. This could involve scenarios like analysing website performance for a specific hotel (e.g. a 5-star London hotel in Q1) and providing a holistic view of the market. For example, AI might identify an increase in average transaction values, potentially driven by a higher industry-wide ADR (average daily rate) coinciding with a major event like a Taylor Swift concert. Additionally, it could notify the hotel of upcoming concert dates in Q4, allowing them to adjust pricing strategies accordingly. We would add additional, bespoke, human insight of course, but by automating a large amount of data analysis, we aim to empower hotels with data-driven insights for strategic marketing decisions.

How can AI support the creation of target audiences and personas to enhance personalisation and build bespoke experiences for brands?

We work with a variety of luxury hospitality and travel brands to develop personalised customer journeys and enhance digital marketing performance by defining target audiences and creating detailed personas. However, building these personas and audiences manually can be time-intensive and often requires extensive analysis of customer data and behaviours, which vary widely across brands. We believe that AI has the potential to streamline this process by analysing vast datasets from sources like Google Analytics, social media, and CRM systems to identify patterns and traits. By exploring AI-driven segmentation and persona generation, we hope to enable brands to offer more tailored experiences, with the flexibility to refine and adapt personas over time as audience behaviours evolve. Our goal is to understand the feasibility of using AI to make audience creation more efficient, scalable, and adaptable, while maintaining accuracy and relevance to drive customisation.

How might we design a system that automates and streamlines communication between students, maintenance teams, and accommodation services for repairs reporting, while reducing manual intervention and improving service delivery?

Background:
• The University of Edinburgh Accommodation Services manages residential facilities for over 9,000 students
• Current challenges:
o Multiple disconnected platforms for maintenance and communication
o Heavy dependency on human input for processing and routing repair requests
o Lack of integrated communication flow between stakeholders
o Manual intervention required at multiple stages of the repair process
• Key stakeholders:
o Property and Residential Services
o Estates Operations
o Students/residents
o Maintenance teams
• Impact potential:
o Improved response times
o Reduced administrative burden
o Better tracking and reporting capabilities
o Enhanced student experience

How might we create an intelligent system for managing large group accommodation bookings that standardises data input, automates processing, and eliminates manual reformatting while maintaining accuracy for large group bookings (up to a maximum of 500+ guests)?

* Current process challenges:
* Reliance on manual spreadsheet management
* Inconsistent data formats from clients
* Time-consuming manual data entry into database
* High administrative burden for large group bookings (500+ guests)
* Key considerations:
* Multiple data points per guest
* Various booking periods
* Different accommodation types
* Special requirements and group arrangements
* Impact potential:
* Reduced administrative time
* Improved data accuracy
* Streamlined booking process
* Enhanced service efficiency

Can pricing of holiday rental accommodation be automated?

We run Bookster, a Property Management System for holiday rentals. The ‘Dynamic Pricing’ tools that we provide get a lot of push back from Property Managers… is there a better way?

Using 2022-24 book festival data, can you help us to identify optimal performance days and times for Full Time, Under 26 and Student bookers to help us better schedule future Festivals and ensure that appropriate slots are protected for the purpose of audience development.

https://www.edbookfest.co.uk/
Some more context around our broader interests in the ticketing data.
. Transaction patterns – how many and which types of tickets are more commonly purchased when (in advance vs near future vs last minutes). We can use this information to inform our marketing campaign planning for 2025.
2. Kids event group patterns – what are the most common combinations of ticket groupings for kids events (including, if possible, against times of events). We can use this information to ensure that we are setting our individual event ticket limits at an appropriate level for 2025, and better understand where group tickets with a single price might help us drive income further into the future.
3. Attendance patterns – how many of what types of tickets are attending when, and are there events which buck the general trend (ideally, also applying this to postcodes so we can understand when non EH bookers attend)? We can use this information to inform event scheduling in 2025.
4. Postcode analysis – mapping different ticket concession types against postcode maps so we can understand where different types of bookers index most highly. This will help inform our advertising media placement for 2025.

How can the Edinburgh International Festival (EIF) effectively identify and engage future supporters—those likely to become dedicated members and donors—by leveraging data insights to target the right individuals at the right time?

Each year, EIF draws a varied audience, from first-time attendees to long-standing patrons. While some of these attendees go on to become regular supporters through membership or donations, the Festival currently lacks a clear, data-driven approach to predict which individuals are most likely to make this transition. Without this clarity, it’s challenging to strategically nurture and engage audience members with high potential to deepen their commitment and support.

The Festival possesses historical data dating back to 2014, encompassing ticket purchases, attendance patterns, event recency, and donation history. Analysing this information could help identify trends and behaviours that signal when an attendee is likely to become a supporter or increase their membership level. For example, patterns may emerge showing that, after attending for a certain number of years or making a specific number of purchases, individuals are more likely to contribute financially or become members.

Armed with these insights, EIF could develop a more targeted and meaningful approach to audience engagement. Instead of broad, untargeted outreach, the Festival could implement personalised campaigns aligned with each person’s level of engagement. For instance, they could encourage donations from loyal attendees who haven’t yet donated or promote membership upgrades for frequent participants. By aligning their strategies with audience behaviour, EIF could foster stronger connections with attendees, enhance funding opportunities, and provide a more personalised experience, all while maximising the impact of their outreach efforts.

What collaborative approaches can stakeholders in the visitor economy, including residents and communities, adopt to address immediate impacts while building a sustainable and livable Edinburgh for the future?

Please refer to the attached document for the background information which includes links to other relevant documents, (EVEP AIMday Question 1).

How can satellite imagery and remote sensing technologies be used to assess the economic, social and environmental impact of Edinburgh’s visitor economy by monitoring tourism-related business activities, visitor patterns, and environmental effects, to inform better decision-making and city planning?

Please refer to the attached document for the background information, (EVEP AIMday Question 2)

How might we identify and validate the optimal design features for an AI-powered tourism itinerary builder that enables meaningful dialogue, considers group dynamics, and creates choice architectures that encourage neighborhood exploration, while ensuring accessibility, sustainability, and ethical AI principles?

Forever Edinburgh, in partnership with Whereverly, has demonstrated the potential of AI-powered tourism assistance through their successful chatbot implementation. As this technology evolves, there is a critical need to understand the design principles that would make such systems more effective and inclusive. Key design considerations include: how to structure AI dialogue for both individual and group itinerary planning, which delivery mechanisms (WhatsApp, web, apps) best serve diverse user needs, how to present choice architectures that balance guidance with user autonomy, and what infrastructure design would ensure security while minimizing environmental impact. This design research challenge seeks to explore how these elements could work together to create more personalized, sustainable, and equitable tourism experiences that encourage exploration beyond traditional hotspots. The insights generated will inform future developments of Edinburgh’s tourism tools, contributing to a more distributed and sustainable visitor economy across the city’s neighborhoods.

https://edinburgh.org/

We seek innovative approaches to identify and validate new commercial applications for NCAP’s 26 million historical aerial photographs, focusing on untapped markets beyond risk management, while maintaining appropriate ethical use of sensitive historical records.

The National Collection of Aerial Photography (NCAP) holds over 26 million aerial photographs documenting historical events and places worldwide, making it one of the world’s largest collections of its kind. While NCAP successfully serves the European bomb disposal and risk management sector, which currently forms its core business, there is significant untapped potential in its vast archives. As a public institution funded through commercial enterprise, NCAP must balance its mission of preserving and providing access to these historically significant records with the need to generate sustainable revenue. The collection, which spans military reconnaissance, urban development, environmental change, and cultural heritage, potentially holds value for diverse sectors beyond its traditional market. However, developing new commercial applications requires careful consideration of ethical implications, data sensitivity, and the preservation of the collection’s historical integrity.

Can the collection of our Scottish visitor management data (based on a small number of metrics and based on a Red, Amber, Green rating) which is collected by various organisations between April and October (annually) and then collated in Excel be automated in some way to make this process easier and more efficient, and could the resulting data then be visualised in a more effective way, for more impactful and effective reporting to Government?

This visitor management data has two parts:
• RAG Rating (Red, Amber, Green)
• Qualitative data collection based on limited number of metrics – ie how many positive visitor interactions, negative interactions, bags of rubbish collected, fires, etc. Data show’s the organisation’s capacity to deal with a number of visitor management issues.
Each organisation such as national parks, councils, etc, collects data on a regular basis between April and September each year, they either complete a form, put it in Excel, or send an email to the point of contact at NatureScot. The results are then collated and put into an Excel spreadsheet by the point of contact who manages the collation and reporting.
Observations (2024) – most teams have reported reduced capacity to collect the data in the first place because less funding/resource for rangers on the ground who can collect this data. However, gaps in the data also make the reporting appear as if everything is being managed successfully when this isn’t necessarily the case.

How can today’s creative digital technology interpret the disappearing industrial archaeology of the Wanlock Valley, Dumfries and Galloway?
Can such an interpretation:
i)act as an educational tool to support the National Curriculum for Excellence and post-graduate study, and
ii) attract a greater number of visitors with broader interests to enjoy the sense of place and benefit from broader storytelling of this important history and by so doing increase the resilience and sustainability of the museum?

The Museum of Lead Mining (MoLM), Wanlockhead, is documented on the website www.leadminingmuseum.co.uk and Available Guides | Bloomberg Connects. It tells the story of this community and represents the once important social and lead mining industrial history between1700 and 1950.
The obstacle
MoLM is a small, independent museum run by the not-for-profit charity Wanlockhead Museum Trust, SC001508.
To be resilient and sustainable we need to double our footfall from ~5000 to 10,000 p.a.
Is it worthwhile?
The museum preserves and conserves the artefacts related to this history; promotes the local history; educates the general population. It is unique in Scotland, especially as members of the public can enter Lochnell mine and experience the old way of life.
The museum is accredited by the Arts Council of England and holds a collection recognized by the Scottish Government, through Museums Galleries Scotland, as being of national historic importance.
Wanlockhead village is a conservation area. The Wanlock Valley has Scheduled Monument status and is a SSSI.
Is it timely?
WMT has been following a path of continuous improvement. Recent years have focused on securing the foundations of the museum through renovation of buildings and reinterpreting displays. The Board is now looking to future strategic development and this potential project is perfectly timed to lay the foundation for industrial archaeological regeneration projects under consideration.
Is it good for problem solvers?
The Board has an open mind about development and this has led to creative ways of making the museum resilient through the last few, challenging years. A skills audit has highlighted a lack of IT skills to enable future development.
The challenge for external problem solvers is how to utilize modern digital representation to uncover and reinterpret the story, both with surface engineering and the underground workings. This lead mining complex was one of the early renewable energy sites harvesting and utilizing the significant rainfall to power the pumps to keep the mines drained. We have a significant material in photographs, engineering drawings and old maps as well as a working relationship with the archivist of the land owner, and our patron, the Duke of Buccleuch and Queensberry.
Is it good for funders?
We have already had early discussions about regeneration and reinterpretation of the area with funders such as Historic Environment Scotland, D&G Council – Tourism and the South of Scotland Destination Alliance.
Digital reinterpretation of this history would be an excellent first step towards making the case for regeneration and conservation funding which in turn would attract a greater number of visitors to the South of Scotland.

With the traveltech landscape becoming more integrated, how can we continue to experiment without risking an inconsistent customer experience across platforms and providers?

Industry trend toward consolidation in travel booking platforms
Users commonly research across multiple platforms before booking
Meta-search platforms aggregate experiences from various providers
Multiple touchpoints exist: mobile apps, websites, partner integrations
Competitive market driving need for rapid innovation
Complex customer journey spans different platforms and devices
Growing expectation for seamless cross-platform experience

How might we encourage travelers to adopt our app given that travel is fairly episodic with a longer behavioural cycle (compared to say food delivery, rideshare, etc.)?

Travel planning happens infrequently compared to daily-use apps
Long gap between research phase and actual booking
Higher customer acquisition costs due to infrequent usage
Challenge of maintaining engagement between trips
Competition from super-apps offering multiple services
Seasonal nature of travel affecting usage patterns
Higher transaction values but lower frequency of use

With the rise of AI, how can the travel industry continue to build trust/authenticity into their products?

Reviews and recommendations crucial for booking decisions
Growing awareness of AI-generated content in travel
Authenticity of local experiences remains key selling point
Travel purchases involve significant financial commitment
Increasing use of AI in customer service and recommendations
Importance of cultural understanding and local knowledge
User-generated content serves as primary trust signal
Rising demand for personalized travel recommendations
Need to balance automation with human expertise

How can we model audience similarities to provide personalized recommendations in travel and hospitality based on user data?

Most online travel agencies prioritize “best sellers” and generic product recommendations, often without considering users’ personal information. We aim to test whether we can better match audiences by recommending specific activities based on live data, such as the activity’s finishing location, user details (e.g., a booking for two adults on Valentine’s Day might suggest a romantic relationship, while two adults and two children would call for family-friendly suggestions), and relevant topics (e.g., if the activity is a Harry Potter tour, we could recommend a bar that serves butterbeer or has a related theme). We would like to see if there is an increase in how much travellers enjoy their experience and in future conversions.