Dynamic World is a new landcover product developed by Google and World Resources Institute (WRI). It is a unique dataset that is designed to make it easy for users to develop locally relevant landcover classification easily. Contrary to other landcover products which try to classify the pixels into a single class – the Dynamic World (DW) model gives you the probability of the pixel belonging to each of the 9 different landcover classes. The full dataset contains the DW class probabilities for every Sentinel-2 scene since 2015 having <35% cloud-cover. It is also updated continuously with detections from new Sentinel-2 scenes as soon as they are available. This makes DW ideal for change detection and monitoring applications.
A key fact about this dataset is that Dynamic World is not a ready-to-use landcover product. Users are expected to fine-tune the output of DW with local knowledge into a final landcover product. Since DW provides per-pixel probabilities generated by a Fully Convolutional Neural Network (FCNN) model, a lot of difficult problems encountered in classifying remotely sensed imagery are addressed already and allows users to refine it with a relatively simple model (such as Random Forest) with small amount of local training data.
A good mental model to use for Dynamic World is to not think of it as landcover product but as a dataset that provides 9 additional bands of landcover related information for each Sentinel-2 image that can be refined to build a locally relevant classification or change detection model.
As seen in the mangrove classification example, using the Dynamic World probability bands as input to a supervised classification model can help you generate a more accurate landcover map in less amount of time. It also eliminates the need for post-processing the results.
To test this concept and explore the potential of this new dataset in developing locally relevant landcover maps – I partnered with Google and WRI to develop a training workshop and host a 5-day “Mapathon” with participants of diverse backgrounds. The event was a mix of hands-on workshop along with hackathon-style group projects to use Dynamic World for a real-world application.
The workshop was hosted by Regional Centre for Mapping of Resources for Development (RCMRD) in Nairobi, Kenya. You can read more about the event in this article. I and Elise Mazur from WRI also gave a talk about our experience at Geo for Good 2023.
In this post, I want to share more technical details about the workshop materials and code for projects for those who may want to use Dynamic World for their own applications.
A long awaited community event that happened after a break of 3 years. The conference was attended by over 200 people from across the world. I got a chance to participate and meet many of QGIS community members in-person.
In this post, I hope to share some of my insights and resources from the conference.
This was the meeting point of over 100 QGIS developers, users, and trainers from all over the globe. It was the first time I met the QGIS community in person, including some of the people whose work I have admired for years. The event took place over 3 days – 2 days of workshops and 1 day of talks. I am putting some of my notes, takeways and links to resources shared on other channels (twitter, telegram, email) for the benefit of folks who were not present
I recently attended the OpenAQ workshop in Delhi . The workshop’s goal was to bring tech, science and media folks working on air quality together and brainstorm how to use open data to tackle air pollution challenges. Below are my notes and links to materials presented during the workshop.
Below are my notes and links to materials presented during the workshop.
I was at #cartonama conference in Bangalore on 23rd September, 2012. It was wonderful to be among geogeeks and see the energy and enthusiasm. Got to see some great demos and LBS apps from startups in India. Here’s a run down from my notes and bookmarks.
Traffic cams in Bangalore: This was probably the coolest demo for me. Love the simplicity and usefulness. Just open this URL in your smartphone browser and get real-time feed from the traffic-cam neareast to you from BTIS.http://hackday.freecrow.org/trafficam
Thru the Gall: A Layar-based android app to solve the last-mile direcction problem in India. Bubbles guide you through the narrow-lanes (galli). http://www.layar.com/layers/galli/
Padma: Search and Browse Indian documentaries through locations of the clips in the documentaries http://pad.ma/map
DelightCircle: LBS app to discover offers and discounts near you. Surprised at the coverage and partnerships they are able to genrate in the short time. Powered by MongoDB at the backend.http://delightcircle.com/
Commonfloor.com: Google Maps API app for property search. Nicely done and simple UI and fast lad times. The team talked about challenges and techniques they use to de-dup the listings and pin the individual listings to the apartment block. http://www.commonfloor.com/apartments-for-sale/maps/
Yahoo: I had forgotten about yahoo as a geo player. But from the talk the team seems pretty excited about their geo APIs. The Geo APIs are used by many yahoo properties internally and available to developers too. The API to extract locations from free-form-text was cool. YQL is a SQL-like query interface for their API. http://developer.yahoo.com/yql/console/, APIS at http://developer.yahoo.com/geo/
NextDrop: Very interesting idea. Most places in India do not have a regular water supply and people’s day revolves around when they are able to collecct water. They have built a system that alerts subscribers by sms when they can expect water based on the information from the utility company. Subscribers pay Rs.10/month to get this data. Their challenge revoled around locating their customers on a map – both for billing as well as determining which pipeline will supply them water. http://nextdrop.org
Lokasi: Simple android app that allows you to share your current location by sms. Reverse-geocoding + location identifier. http://www.onze.in/ . collected POIs + street geometry for entire bangalore is about 6 months with a 10-person team.
India’s first FOSS4G – India conference was organized on 25-26 October, 2012 at Hyderabad. This was a surprisingly low-profile event and there was very little buzz about it on the internet. I learnt about it just in time to register and make it to the event. Here I am sharing some notes and impressions. (notes are mostly from memory, so please excuse if I missed something)
I haven’t been to other FOSS4G events, but this was unlike any other Geo conference I have attended. It was very formal and government-centric. All presentations were from government researchers, employees or students.
The penetration of open-source software within government was deeper than I anticipated and the decision-makers – including the politicians recognize the benefits of using open technologies. The Chief Guest – Shri V. Aruna Kumar, MP from Rajahmundry district commented that he realizes that the real cost of software is not the upfront cost, but the lifetime maintenance cost and open-source and open-standards can really benefit in the longer run.
Keynote’s main message was that we should really focus on delivering applications that solve real problems. According to the speaker – Shri Anoop Singh, Special Secretary to AP Government – we have not even begun to realize the full potential of GIS within governments. For most end-users and decision makers, the ‘Science’ or ‘System’ in GIS isn’t important and GIS should really be about Geographic Information Services.
India is seen largely as a follower in the open source movement and no big open-source projects have come out of research institutions here. To address this, International Institute of Information Technology released one of their projects – LSIViewer as open-source. The github upload was done during the inauguration of the conference by the chief guest!
Prof. Venkatesh Raghavan receiving 2012 Sol Katz award Prof. Venkatesh Raghavan, Osaka City University was chosen as the winner of the 2012 Sol Katz ward by OsGeo. The award was given to him during the conference.
Postgrasql + PostGIS, Geoserver and Openlayers seem to be the platform of choice. Almost all applications I saw were built using one of these.
ELOGeo portal was shown an example of learning resources available for open source geo tools.
There is virtually no open data for India given the strict geo data policy. But even within the government departments – there is a lot of reluctance in data sharing. Everyone talked about this problem of data being guarded very closely and seldom shared. But some forward thinking departments, especially in Tamil Nadu, have found a way around it by promoting Web Services. They publish a WMS feed for the datasets. That way other departments can use the data without really having any ownership of it. This seems to be working well.
Ms. Mahalakshmi from National Informatics Center (NIC) in Tamil Nadu demonstrated some of most interesting internal applications built by them. They have built web services for different departments and each department uses these web services for their application. Administrative boundaries, census data, land use maps, land parcels, police data and planning related data are some examples which are available via WMS. She also showed a pretty cool app that BSNL – Chennai uses to identify fault points in their 2G network. The map is updated in real-time with pings from their different 2G sites around Chennai.
BHUVAN – Indian Space Research Organization’s (ISRO) portal for disseminating geodata has come a long way since it’s launch a few years back. It’s built completely using open-source stack and it’s a pretty snappy app. For the first time, a lot of thematic datasets are available freely to the public via WMS. They are also making some medium and low resolution raw/derived data from Indian remote sensing satellites available for download. The downloads are for personal use only, but this is a big step forward in India’s data access policy which has been very restrictive so far.
In summary, some positives and challenges for Open-source software described by various speakers.
Why open source is very attractive to governments
No upfront cost. Don’t have to navigate complicated and time-consuming approval and procurement process. Easy to get started at low-cost and do a proof-of-concept.
Standards Compliant. Rising awareness within the government of being standards compliant. Choosing a proprietary solution which is not standards compliant can cause problems and raise questions from others in future.
No vendor lock-in. Can hire another consultant to run or develop services since all code is available.
Challenges to wider open source adoption
Lack of support. Not many companies providing development and support. Most applications shown were written and maintained by in-house staff. Big opportunities for consultants who can provide support for open source tools.
Change aversion: On the desktop, most people are used to proprietary software from their education or prior work. One example given was reluctance in the government to switch to Open Office. Free versions of MS Office is easy to come by and no incentive to learn and adopt open source solution. Using open-source software in education is a way to increase the adoption.
Lack of Mature solutions: Politicians and administrators felt they had limited time in their tenure to implement whatever changes they want and show progress. There is always the attraction to choose a well known solution and get the results. The key message was that they feel the open-source software is mature but there are no mature solutions based on these that the governments can adopt without much risk.