Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: A Twitter bot I made that posts real-time deforestation data (twitter.com/forestswar)
101 points by imnotanerd on Aug 22, 2020 | hide | past | favorite | 34 comments



Heads up:

[1] This is a project I've been working in about a year now and is an honorary successor to a similar project a friend started.

You can read in detail how it works at https://imnotanerd.herokuapp.com/3.

Basically it takes data from the API at globalforestwatch.org that uses satellite imagery to gather area of potential forest loss and paints it in a map to make it easy to imagine.

There is no consideration about forest regrowth. This bot does not track forestal area changes, it tracks deforestation.

[2] I'd like to get some community feedback. The bot has now been running for quite some time and it looks like now it works with consistency (The GFW API data is not reliable when recently published).

Mainly I'd like to see how this bot can be improved regarding the data treatment and the display of it, as I still think the maps, despite being accurate, are a little "underperformant" and could be improved.

[3] Yes. I posted this twice, I forgot to put "Show HN" in the title the first time and I realised too late. Now I can't edit nor remove the previous submission.


I see a number of issues:

1. no deforestation before July 9th.

2. In 10 days between Aug4 and Aug14 2% of the area was deforested, signifying all forrest will be gone in about 1.5 years. However forrest area in 2015 was almost the same size as reported total. Can't be right. Are you mixing ha and km2 by any chance?

3. No tracking between what is part of Forest Management Act, what are is part of designated production forest, and what is illegal deforestation.

4. No data on loss due to natural causes such as wind, fire, and no data on reforestation numbers.


1. That's the date the bot started tracking. Actually, deforestation data starts in 2015 but I thought it'd be dishonest of the bot to count in deforestation that happened before the bot release. First drafts of the bot worked this way but that behavior was abandoned.

2. What do you mean by 2%? I guarantee I'm not mixing units, the remaining area shown is the remaining forestal area for the country in display, meaning that if the deforestation rhythm keeps that way, Suriname will be completely deforestated in 1.5 years.

3. The GLAD alerts dataset, which the bot takes as data source, don't make any distinction between forestal area of any type, they just track deforestation based on changes comparison in satellite imagery, Landsat 7 & 8 being the main imagery sources.

4. Then again, GLAD alerts does not provide any info on the causes of the deforestation.

I think globalforestswatch.org have an API based around industrial logging that let's you track industry-based deforestation only, but it's based on in-site reports and it's not as reliable as satellite-based GLAD alerts.


Thanks for the reply. For point 2, the source states that from 2001 to 2019 1.3% of the tree cover was lost [0]. However in the twitter feed the loss is 2% per 10 days.

[0] https://gfw.global/3gm1aRN


You are looking at the forest cover lost in Suriname only for that time period.

The bot gathers global forest cover loss for a given day and compares the aggregated of every country against the selected country from the list.

I thought I made it already pretty clear on the way the statuses are written, how would you write them to make it clear that its global area compared to that country rather as that country area only?


Point 4 might be irrelevant to the cause and more importantly part of a negative (hard to track) feedback loop. A patch of forest destroyed by human activity might be more susceptible to stronger winds or extreme weather in general.


"Here we analyse 35 years’ worth of satellite data and provide a comprehensive record of global land-change dynamics during the period 1982–2016. We show that—contrary to the prevailing view that forest area has declined globally—tree cover has increased by 2.24 million km2 (+7.1% relative to the 1982 level). This overall net gain is the result of a net loss in the tropics being outweighed by a net gain in the extratropics."

- Song, X., Hansen, M.C., Stehman, S.V. et al. Global land change from 1982 to 2016. Nature 560, 639–643 (2018). https://doi.org/10.1038/s41586-018-0411-9

This does not take into account loss of biodiversity or effects on climate, but the viewpoint that we are "running out of trees" is a common one. There are actually more trees on earth than stars in the Milky Way by an order of magnitude.

I would love to have my opinion changed on this however, so if further data contradicts the above study (not that one single study ever proves anything), please make me aware of it.


When I started this project I also had that feeling that we were running out of trees.

But my motivation never was to alarm people that we were running out of them, my intention was to make actual deforestation data accessible and approachable to the public so they could decide how big of a problem this was. I have myself learned that I should probably worry less about it than I used to thanks to this bot, and that's fine.

I also think you are not seeing the big picture here. 30 years is pretty much just the minimum for the average tree to grow to a size where it can foster other biomass. On the bigger picture:

"Forests cover 31 percent of the world’s land surface, just over 4 billion hectares. […] This is down from the pre-industrial area of 5.9 billion hectares."

-Earth Policy Institute. http://www.earth-policy.org/indicators/C56/

1.9 billion hectares down in about a century and a half.

My viewpoint now is that there is a deforestation problem in the sense that we destroy forest dependant biomass faster than what it takes for it to recover, but we are still in the time window where the already made harm and the future impact of our deforestation can be reversed, from the viewpoint of Hansen et al we seem to have reversed the negative tendency in tree cover.


Hey, just to clarify this was not a jab at you, I just wanted to present a different viewpoint. Very often people fall into the habit of seeing every environmentalist point as good and true, and anything that contradicts that is false and harmful. A more nuanced perspective is the actual truth.


I didn't take it as jab!

I actually appreciate that you put in effort to add to the discussion and I presume you made some research to find the Hansen 1982-2016 research paper.

That is a success to me because my work has achieved it's goal, discussion about deforestation is moving forwards! It's even more exciting that this is coming from someone who is not a close-minded environmentalist. In a way I agree that, despite taking care of our environment should be a priority, doctrine-like following of mantras and lack of proper criticism is actually harmful for any movement and individual.


Hi Daniel, it's really nice project.

Any plan to build a website for more visualization in the future or maybe a newsletter to receive the updates?


Kind of.

Not exactly a project inside this project, but maybe a separate visualization tool using the GLAD alerts API to allow people to gather the same information the bot does, but in an easy to use website instead of having to use the package itself inside their machines. Unfortunately I'm not much of a designer so this could take a long time.

The newsletter idea sounds cool. Maybe even send detailed reports? There is more data the bot knows on each request but gets ignored for simplicity of the statuses.

However, as of now I'm more concerned about the continuity of the bot, as mLab (where the bot database is) will be terminating their service by November. Currently looking for a free or affordable place to deploy this app.


> The newsletter idea sounds cool. Maybe even send detailed reports? There is more data the bot knows on each request but gets ignored for simplicity of the statuses.

Indeed a detailed report would be nice to receive.

About mLab, I think you could give a shot to the MongoDB Atlas Free Tier (M0).


This seems cool, but somewhat unconstructive. Feels like receiving daily tweets about the continuous destruction of wildlife (but being able to do nothing about it personally) is a very negative overall experience, and doesn’t seem healthy. It is in effect receiving daily bad news which you can do nothing about.


I personally know many people who are now doing whatever little they can only because they got exposed to ongoing wildlife catastrophe through social media. Yes, they may not make a difference on a global scale considering them in isolation, but there are many more such people and the effects add up perhaps to make a minor but nevertheless significant difference in the long term.

Hiding the information away only helps the exploiters to win decisively.

Yeah the data is somewhat sensationalist, but still it is not false data and many people who would otherwise not bat an eye can sometimes be impacted into change by sensationalist or shocking presentation of data.


I dunno, speaking personally, the world has so many problems that it's best (for me) to not think that much about them, or be exposed to that much information about them. Maybe this is a bad attitude, but ignorance is bliss.

If I was like Bill Gates or something, I think I would have a different attitude. Only when information can actually inform my decisions does it become meaningful. No matter what I hear about global deforestation, I'm not going to do anything about it. So why even hear it in the first place? It just serves to pollute my mind and make me feel bad.


> the world has so many problems that it's best (for me) to not think that much about them, or be exposed to that much information about them

I think a lot of people feel that way, and it's not an entirely irrational response.

Perhaps the best solution is to choose just one or two problems to worry about and try to help solve. Let other people worry about the others.

The "Think Globally, Acr Locally" mantra suggest it's easier to influence your local community and government, rather than a global, hard to wrap your head around, abstract issue.

Some people may still decide that working on global deforestation is the most important thing to do.


Not unconstructive at all, just uncomfortable. Awareness to what is happening to our planet is key!

Taking down forests is unconstructive. :)


No you can do things about it. If more people are talking about deforestation, political parties realize it becomes a way to swing voters.


To talk about the topic, one must first understand about a topic. Citing a twitter number does not help when it's not even known how much was destroyed by winds or fire and how much is reforested. Remember, wood is a renewable source, and chopping down trees is part of healthy forest management.


> it's not even known how much was destroyed by winds or fire and how much is reforested

I know that's an important design flaw: it can't differentiate between man-made deforestation and natural forest change. I'd like for this bot to not exist until there is a way to reliably measure the former.

But if I waited for that to happen probably I'd never have the chance to understand myself and make others see the problem being discussed here.

Moreover, reforestation is not taken in consideration because there is no evidence that reforestation undoes the damage causes by deforestation. Yes, the tree-covered area is there again, but is the biomass, diversity and ecosystems that the disappeared trees fostered there again? That's the real issue here. Let alone the fact that in most reforestation it takes decades on average for the trees to simply grow.

Wood is a renewable resource as long as the rhythm at which we exploit it it's sustainable.


Are you saying we shouldn’t report bad news if we can‘t do anything to help?


It's called awareness raising.


Yes. For an even more extreme experience try following @AuschwitzMuseum. See how it affects your daily life and priorities. Does it make you cry sometimes? Lucky you. There is a very beautiful and important need called mourning. Enjoy.


I somewhat understand your point here. In fact this bot has sometimes made feel a little bit miserable.

I detailed this aspect of the bot in a blog post (imnotanerd.herokuapp.com/4).

TLDR: I think this bot remains important because it makes true deforestation data accessible for whoever it may concern and it helps people understand and track the issue in a way that provides certain "feedback" against traditional absolute deforestation data like "x football pitches per minute".


It's more negative for the flora and fauna I would think.


Why does it only tweet about Suriname?


It will make map comparisons with Suriname until this country is deforestated, then it will move to the next country in the list.


What’s the Suriname for?


Just happened to be the first country on the list.


How can it be real-time?


A little bit of marketing wording I must confess, but it is as real time as you can get with this kind of datasets.

The API that I used for this project along the other APIs by GFW power an app called Forest Watcher that it's pretty real-time.


So if you found a security bug in some networks software you'd just make it public? zero validation? just Tweet it out publicly as soon as you found it?

What due diligence did you do on globalforestwatch.org?


I emailed them about each and every flaw of their API that I detailed, and some other details I didn't reveal about it.

Their developers and public relationships people are good professionals but I can't stop feeling they are overworked.

Moreover, I detailed those flaws not as empty criticism but rather as instances of software being problematic not because of undocumented behavior but actually because of "undocumented design".




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: