Photo by Devam Jhabak on Unsplash
Social media has been significantly impact on customer buying decisions in many business sections. Online travel agencies (OTAs) are not the exception. They are facing the challenge of incorporating new media channels into their existing value chain while optimizing online conversions and online booking experience. To address the above challenges, an end-to-end BI use case was developed. The process encompassed concept creation, database design, building data pipelines, and dashboard production. A traveller insights dashboard was developed to present four KPIs, daily metrics, and search trends, which can be used to support travel product marketing and optimise OTAs conversion rates throughout the entire traveller journey.
Social media, such as Facebook, Twitters, YouTube, Instagram, and the TikTok, have gained importance and have dramatically influenced everyone's behaviour when it comes to decision-making. This phenomenon is even more noticeable in Tourism. The increase in popularity of Instagram changed people's search path from search engines.
Websites that provide travel-related products have a 98% abandonment rate among all businesses. Statista (2022)
Online travel agencies (OTAs) face low conversion rates due to high shopping cart abandonment. The complexity of the booking process and abundance of options lead to longer decision-making times for consumers. According to Expedia (2015), travellers visit an average of 38 websites before booking a trip. This is because the booking process involves multiple steps and can be stressful due to the massive products and options. This can delay the traveller from making purchases.
Social media has been significantly impact on customer buying decisions in many business sections. Online travel agencies (OTAs) are not the exception. To captures critical touchpoints along a tourist's decision-making phases, while providing valuable insights for business providers, a traveler insights dashboard we developed for the Swiss OTA offers real-time information from Instagram Posts and Google Trends. It enables travel service providers to enhance the shopping experience by leveraging social media and search engine data to understand travelers' activities, preferences, and online conversion.
Having a data collection strategy is important for every business intelligence project. To do so, I mapped out traveller's digital journey, and list possible data touch points that can trigger traveler to visit a OTA website, and make a purchase to post on social media to share their travel journey.
When mapping out a traveller's digital journey, we discovered that social media has gained significance in their decision-making process (Smith & Anderson, 2018). Instagram travel feeds have become powerful indicators that influence travellers' choices in planning upcoming trips (Han & Chen, 2021).
For the data collection, I used the most influential data sources across traveller digital touchpoints ─ Google trend and Instagram Hashtag matric to draft the data visualisation prototype.
To better understand the impact of social media and online search on travelers' buying decisions, and to assist Swiss online travel agencies (OTAs) in improving their online booking experience and marketing effectiveness, I used Figma to make a mockup to visualise key metrics and trends by analyzing social media and search engine data that will be show on the prototype.
We used Redshift t. I used Airflow to automat ETL and ELT workflow to fetch, transform and analyse a total 10K IG posts and Google trends
To build a database is built using the following data engineering techniques
We utilized data pipelines to address critical business questions and present key metrics in our proposed travel insights dashboard. The IG Hashtag Statistic graph provides information on hashtag usage in Instagram posts, helping us to achieve business goal:
Questions 1: “How can we improve the tourist travel experience in Switzerland by leveraging social media and search engine data? “
Questions 2: “How do individuals and businesses use social media hashtags to enhance their online presence and performance?”
Based on the data analysis, posts with the hashtag Zurich received the most likes, with 169,235 likes and 6,170 comments. Valais and Luzern were the second and third most popular places, with 144,957 and 120,267 likes respectively. The doughnut chart also revealed that Zurich, Wallis, and Uri were the most popular cantons, and Zurich, Zermatt, and Interlaken were the most admired cities. Therefore, we can conclude that Zurich is the most popular and most visited place in Switzerland. The full dashboard can be viewed in Tableau Public
The end-to-end process of constructing a database, from data collection to data visualization, has enabled us to generate valuable traveler insights and answer our business questions. The proposed dashboard is designed to create value for business providers, including online travel agents and marketing firms, by offering key metrics that can be utilized to optimize their services and products, and provide travelers with more personalized solutions.
Initially, our plan was to obtain location tags from each Instagram post to create an Instagrammable Spot Hunting app. However, Instagram deprecating the local tag endpoint from its API, we had to change the plan by using the hashtag #inLoveWithSwitzerland, a hashtag that is used by the Swiss national tourist organization - MySwitzerland. Moreover, the English names of the 26 cantons in Switzerland are also deployed as the starting point to fetch relevant hashtags from Instagram posts. Although we couldn't capture all the locations in Switzerland, the information we collected was sufficient to create a minimal viable prototype for this BI project.
It was a great learning experience because it has allowed me to learn many hard skills in data engineering. I have acquired fundamental skills and knowledge in interacting with databases and different applications, as well as the know-how for combining the Design Thinking method into product development. Furthermore, the opportunity to work with popular software such as AWS and Airflow has been invaluable in bringing the business idea to life and creating a magnetizable product.
Airflow Dag:
Jupyter Notebook:
