After a strenuous year fighting a lawsuit against the company she co-founded, Tinder co-founder Whitney Wolfe Herd joined Russian entrepreneur and Badoo founder Andrey Andreev to launch Bumble....

Foreword

This dataset is a preview of a bigger dataset. My Telegram bot will answer your queries for more data and also allow you to contact me.

Context

When Dating apps like Tinder were becoming viral, people wanted to have the best profile in order to get more matches and more potential encounters. Unlike other previous dating platforms, those new ones emphasized on the mutuality of attraction before allowing any two people to get in touch and chat. This made it all the more important to create the best profile in order to get the best first impression.

Parallel to that, we Humans have always been in awe before charismatic and inspiring people. The more charismatic people tend to be followed and listened to by more people. Through their metrics such as the number of friends/followers, social networks give some ways of "measuring" the potential charisma of some people.

In regard to all that, one can then think: - what makes a great user profile ? - how to make the best first impression in order to get more matches (and ultimately find love, or new friendships) ? - what makes a person charismatic ? - how do charismatic people present themselves ?

In order to try and understand those different social questions, I decided to create a dataset of user profile informations using the social network Lovoo when it came out. By using different methodologies, I was able to gather user profile data, as well as some usually unavailable metrics (such as the number of profile visits).

Content

The dataset contains user profile infos of users of the website Lovoo.

The dataset was gathered during spring 2015 (april, may). At that time, Lovoo was expanding in european countries (among others), while Tinder was trending both in America and in Europe. At that time the iOS version of the Lovoo app was in version 3.

Accessory image data

The dataset references pictures (field pictureId) of user profiles. These pictures are also available for a fraction of users but have not been uploaded and should be asked separately.

The idea when gathering the profile pictures was to determine whether some correlations could be identified between a profile picture and the reputation or success of a given profile. Since first impression matters, a sound hypothesis to make is that the profile picture might have a great influence on the number of profile visits, matches and so on. Do not forget that only a fraction of a user's profile is seen when browsing through a list of users.

https://s1.dmcdn.net/v/BnWkG1M7WuJDq2PKP/x480" alt="App preview of browsing profiles">

Details about collection methodology

In order to gather the data, I developed a set of tools that would save the data while browsing through profiles and doing searches. Because of this approach (and the constraints that forced me to develop this approach) I could only gather user profiles that were recommended by Lovoo's algorithm for 2 profiles I created for this purpose occasion (male, open to friends & chats & dates). That is why there are only female users in the dataset. Another work could be done to fetch similar data for both genders or other age ranges.

Regarding the number of user profiles It turned out that the recommendation algorithm always seemed to output the same set of user profiles. This meant Lovoo's algorithm was probably heavily relying on settings like location (to recommend more people nearby than people in different places or countries) and maybe cookies. This diminished the number of different user profiles that would be presented and included in the dataset.

Inspiration

As mentioned in the introduction, there are a lot of questions we can answer using a dataset such as this one. Some questions are related to - popularity, charisma - census and demographic studies. - Statistics about the interest of people joining dating apps (making friends, finding someone to date, finding true love, ...). - Detecting influencers / potential influencers and studying them

Previously mentioned: - what makes a great user profile ? - how to make the best first impression in order to get more matches (and ultimately find love, or new friendships) ? - what makes a person charismatic ? - how do charismatic people present themselves ?

Other works: - A starter analysis is available on my data.world account, made using a SQL query. Another file has been created through that mean on the dataset page. - The kaggle version of the dataset might contain a starter kernel.

Total-Revenue Time Series for Bumble Inc. Bumble Inc. provides online dating and social networking applications in North America, Europe, internationally. It owns and operates websites and applications that offers subscription and in-app purchases of products. The company operates apps, including Bumble, a dating app built with women at the center, where women make the first move; Badoo, the web and mobile free-to-use dating app; Bumble BFF and Bumble Bizz Modes that have a format similar to the date mode requiring users to set up profiles and matching users through yes and no votes, similar to the dating platform; and Bumble for Friends, a friendship app where people in all stages of life can meet people nearby and create meaningful platonic connections, as well as Geneva app where users can create and join chat, forum, audio, video, and broadcast rooms. The company was founded in 2020 in and is headquartered in Austin, Texas.

Change-To-Inventory Time Series for Bumble Inc. Bumble Inc. provides online dating and social networking applications in North America, Europe, internationally. It owns and operates websites and applications that offers subscription and in-app purchases of products. The company operates apps, including Bumble, a dating app built with women at the center, where women make the first move; Badoo, the web and mobile free-to-use dating app; Bumble BFF and Bumble Bizz Modes that have a format similar to the date mode requiring users to set up profiles and matching users through yes and no votes, similar to the dating platform; and Bumble for Friends, a friendship app where people in all stages of life can meet people nearby and create meaningful platonic connections, as well as Geneva app where users can create and join chat, forum, audio, video, and broadcast rooms. The company was founded in 2020 in and is headquartered in Austin, Texas.

TripDurationsDefaultResults of the Bumble-BEEHAVE simulation described in 1. Individual level comparisonColony_Validation_1_250Results of the Bumble-BEEHAVE simulations presented in 2. Colony level comparison (runs 1 - 250)Colony_Validation_251_7500Results of the Bumble-BEEHAVE simulations presented in 2. Colony level comparison (runs 251 - 7500)Population_ValidationResults of the Bumble-BEEHAVE simulations presented in 3. Population level comparisonSingleHabitatPatchResults of the Bumble-BEEHAVE simulation described in 4. Model applications: Setting/Output 1Baron_500initialQueensResults of the Bumble-BEEHAVE simulation described in 4. Model applications: Setting/Output 2 with 500 initial queenBaron_7500initialQueensResults of the Bumble-BEEHAVE simulation described in 4. Model applications: Setting/Output 2 with 7500 initial queens

This dataset was collected as part of a field experiment in 2016. Corbicular pollen loads of honey bees and bumble bees were collected, along with floral resource availabilty from June through September. The contents of pollen loads were idenfied using visual microscopy to determine the frequency of flower constancy, along with the richness and diversity of pollen collected by bee colonies. In addition, the abundance distribution of flower resources was compared to that of bee collected pollen to examine more preferred and less preferred pollen types for these two bee species throughout the foraging season. Floral abundance was sampled seperately for wildflowers, alfalfa, corn, and soybean, so there will be a seperate dataset for each. Data was processed using R, and an annotated RMD script can be found here https://doi.org/10.5281/zenodo.14738621