Sitting on the edge of the cliff, watching a sunset.
Looking at a mosque in the middle east.
Posing with a manta ray.
If you classified the images above into the gender of the uploader, what would you do?
Our team at Quilt.AI wanted to see if we could identify a traveler’s gender from the nature of their uploads.
Creating a Travel Model
We used visual and textual machine learning to classify and cluster images uploaded by 5000 unique travelers. Through this exercise, we saw two dominant polarities emerging in terms of how travel is represented.
Axis 1: This was based on the action being captured in the image. These actions were either activity-focused or reflective in nature. We saw active experiences such as adventure sports and partying that engaged with the physical world, to more quieter and reflective experiences amidst natural landscapes like beaches, rainforests, and mountains that fed the soul.
Axis 2: This was based on the similarity/distance of the action to the everyday life of the traveler. We called it “familiar” vs. “unfamiliar”- ranging from visiting familiar places and doing familiar things but perhaps more intensely, such as visiting churches, libraries, gym, beaches, to experiencing the more unfamiliar, such as mountaineering, helicopter ride, harbors, and castles.
There were certain semiotic conventions we saw in how travel was represented:
The first convention we noted was around the space of ‘Unfamiliarity’ where the “new” is clustered as follows:
- We saw that men would upload unfamiliar activities, and women would upload unfamiliar places.
- Male experiences of the unfamiliar were often steeped in adventure, while female experiences of religious places range from a pilgrimage to historical monuments.
The second convention we noted was around the space of ‘Activity’ where the action was linked to the benefit:
- Broadly, the activity that women engaged with on holidays was focused on beach holidays, connecting with friends in ritual contexts (such as weddings), or bicycling through the countryside.
- The activities men tried to showcase were more “physical” and were designed to show vacation as a moment of conquest and victory.
The third convention we noted was around the space of ‘Reflection’ where meditation/contemplation also varied by gender:
- We noted that the men were uploading “bonding” related content- hanging around campfires- in order to signal reflection.
- By contrast, the women were uploading learning/experience expansion related content- doing yoga, going to libraries and taking photographs of large vistas- in order to do the same.
The fourth convention we noted was around the colours detected from the images posted by men and women:
- Women uploads had soothing colours like off-white, grey, brown, sienna, brown, and exhibited a significant range and variation: misty rose, sea green, snow, crimson, royal blue, light cyan, plum, peach, and light coral.
- By contrast, male uploads were either very bright (azure, turquoise, coral) or dark (maroon, chocolate).
The fifth convention we noted was around emojis and we found that women were far more experimental in their use of non-verbal description:
- There is a greater sense of playfulness in women’s non-textual communication. While we found women using interesting, lesser-used emojis like 🌊(wave), 🎠(carousel horse), 🦄(unicorn), 🧛(vampire), 🧘(yoga), 🦋(butterfly).
- Most men in our sample tended to use emojis denoting everyday objects like 🍻(beer), 🍔(burger) and the commonly-used face emojis (such as 😍😛😂😎).
Overall, we saw that the male recall of travel was linked to activity and assertion (sports, conquest, luxury), while the female recall of travel was linked to experience/memory that was more reflective (architecture, landscapes). It could well just be that travel Instagram happens to be the space where men and women decide to play up these gender norms in order to “fit in” to the expectations of the audience.
Does advertising play to these norms as well? Can you guess what norms these advertisements follow in terms of whether they show “activity” or “experience”?
Read more stories on Quilt.AI.