Malaysia - iNaturalist World Tour

We’re featuring on Malaysia on the World Tour today.

Here’s the GADM Level 1 and Level 2 places we’re using. Again, my usual questions about whether these administrative units are correct and widely used.

What can we do to improve iNaturalist in Malaysia - in particular to grow the community of regular users (as opposed to people who only participate in large annual events)?

iNaturalist is not translated into Malay. Would this help? Or is English widely spoken enough in the country.

1 Like

How well does the computer vision work in Malaysia? Are there enough active identifiers to make iNaturalist a useful tool for species identification?

Globally, around 40% of verifiable observations need ID/confirmation, but in Malaysia it’s up at 75% of observations at needs ID.

Percent of Observations that are “Needs ID”

Globally Malaysia Difference
Fishes 21% 88% 67%
Chromista 51% 100% 49%
Non-flowering Plants 50% 91% 42%
All Plants 45% 85% 41%
Flowering Plants 43% 82% 39%
All Insects 46% 84% 37%
All 40% 75% 36%
Beetles 50% 84% 33%
Arthropods, excluding
Arachnids and Insects
58% 89% 31%
Other Animals 51% 78% 27%
Mammals 15% 42% 27%
Fungi 70% 95% 25%
Hemiptera 60% 84% 24%
Birds 5% 27% 22%
Lepidoptera 35% 57% 22%
Orthoptera 63% 85% 22%
Miscellaneous Small Things
(Viruses, Bacteria, etc.)
70% 90% 20%
Hymenoptera 58% 75% 17%
Flies 78% 94% 17%
Mollusks 45% 61% 16%
Herps 13% 14% 1%
Arachnids 68% 66% -2%

There are also around 33,000 observations with no label, which I was surprised to see, given the focus on volume with the recent “challenge” events, is not hugely different from the similarly populated state of Texas in the US, which has ~29K unlabeled.

3 Likes

Although English is quite widely spoken, particularly by older generations as a result of Malaysia’s colonial heritage, it has been in and out of favour in the education system over the years since the country gained independence, so, particularly in rural areas, there is also a sizeable proportion of the population who for whom having the site translated into Malay would be a big benefit. I noticed several misidentifications during the CNC that seemed to be the result of a lack of confidence in English (I can’t think of exact ones right now, but along the lines of identifying “Bees” as “Bee Flies”, which is entirely understandable once you think from the perspective of an English language learner).

The climate here also makes it challenging to go out observing in the daytime sun (which is probably why there were so many fish market and aquarium, and shopping mall plant observations during the CNC). Hiking is for the hardy (“Mad Dogs and Englishmen” and all that) – I’ve been searching for walking trails within driving distance of home, and most of these seem to be up mountains. I’m much more of a “nice level walk, please” person these days. I have a stinking, rotting shirt set aside for nature walks that will be drenched through within twenty minutes of getting outside.

The computer vision is improving in some areas, but we rely on a handful of individual experts for species level identifications outside of the most “popular”. For most of us, field guides to birds, butterflies, reptiles and mammals are readily available, but they are lacking in other areas – particularly invertebrates. I have one for dragonflies (which has been helpful), and one for phasmids (which will be helpful if I ever see one), but of the others I have, one is specific to insects in agriculture, so only covers pollinators and pests, and the other is focused on South East Asia in general. In both of these, the number of actual species identified is only a handful, with most photos only identified to family or higher. Susan Myers’ “Wildlife of Southeast Asia” devotes 150 pages to birds, 30 to mammals, 15 to reptiles, and 6 species of frog, then ends with 11 butterflies, 1 moth, 1 dragonfly, 1 damselfly, a tropical carpenter bee, a golden orb-weaver spider, and a final page with a leech, a giant pill millipede and a tarantula (none of them identified to species).

So the bad news is that identifying to species level for the majority of us laypeople, for the majority of our observations, is going to be impossible. The brilliant news is that iNaturalist is probably the best way we have of getting at least some of those observations identified.

Two years ago I knew next to nothing about Malaysian wildlife, despite having lived here for most of the past 20 years – I could name a handful of species. Since joining iNaturalist my knowledge has increased a thousand-fold (I even find myself thinking in Latin nomenclature on occasion).

But, after two years, the level of expertise I’ve arrived at is “aware of my own uncertainty”. I’ve now misidentified enough “species that look exactly like a species that they aren’t” and “individuals that bear no resemblance whatsoever to what I thought that species looked like” to be much more hesitant about offering species level IDs (I’ve just been introduced, for example, to the Three-spotted Grass Yellow, which is only one-spot different from the Common Grass Yellow that I thought it was). I’ve also become more hesitant about using the maps on the species pages to judge the likelihood of a computer vision suggestion being accurate – for some less common species (again I can’t give specific examples) I get the impression that there have been enough computer-vision-aided misidentifications that checking the map to see if that species has previously been observed in Malaysia can be misleading.

4 Likes