Check-in
Agenda
WednesdaySep. 19
-
-
Breakfast
Pavilion
-
Keynote
Welcome and Kickoff
PavilionWelcome to SatSummit by Mapbox & Development Seed. Satellite data is part of everything. It’s a multi-billion dollar industry and its moving incredibly quickly. Everyone in this industry wants to use data to improve the planet. Now more than ever. But that doesn’t just happen. Challenge group to make satellite wicked simple, put information in people’s hands.
- Speakers
- Eric Gundersen
- Ian Schuler
-
Keynote
On the Ground
PavilionPractical applications from people using satellite imagery on the ground for international development and urban management - in depth look at Bogotá and Ho Chi Minh City.
-
Keynote
Government
PavilionA look at how the government serves as a data provider and how it can create impactful open data and the move away from scientific data living in a black hole.
- Speakers
- Rahul Ramachandran
- Tom Lee
-
Keynote
The Providers
PavilionToday’s state of the art in satellites for development and where the field is going. Hear from leaders in AI, and space technology. NASA, ESA, DigitalGlobe, and Planet discuss their current collection, analysis and distribution capabilities.
- Speakers
- Sarah Sarah
- Rhiannan Price
- Kristi Kline
- Christoph Aubrecht
- Andrew Zolli
-
Keynote
The Storytellers
PavilionHearing from storytellers, we’ll learn how they use imagery to inspire and inform.
- Speakers
- Ben Grant
- Andy Revkin
-
Keynote
Our Opportunity
PavilionCurrent state of the industry of EO applications for societal benefit, and a look at the future trends of satellites for global development.
- Speakers
- Danielle Wood
-
Break
Coffee
Pavilion
-
Panel
Eye on Gotham
Hemisphere BWe now have the spatial and temporal resolution to keep up with the cadence of a modern city. These presenters show you new ways to look at our urban landscape.
- Moderator
- Jessica Seddon
- Panelists
- Annemarie Schneider
- Denice Ross
- Kien Vu
- Derek Lieu
-
Panel
Friends with benefits: Gov, Commercial, NGOs
Continental CMasters of bureaucratic ju-jitsu show you how to get bureaucracy out of the way so that we can work together on common goals across government, NGOs, and companies with some of the biggest data footprints.
- Moderator
- John Crowley
- Panelists
- Jamie Favors
- Laura McGorman
- Sarah Henley-Shepard
-
Panel
Artificial Intelligence 101
Hemisphere ALet’s peel away the hype and talk specifics on artificial intelligence for earth observations. What can you expect as a customer of AI? What’s the current state of the art and what can AI tell us? What are the pre-reqs in order to do this well?
- Moderator
- Dalton Lunga
- Panelists
- Drew Bollinger
- Sean Gorman
- Jubal Harpster
- May Casterline
-
Panel
Planetary Scale Analysis
PavilionGovernment and commercial imagery providers are making data available on the cloud in analysis ready formats that support planetary scale analysis, AI, and integration into web apps. Stories from the trenches on how to get data autonomously prepped for analysis and what it means to have your own workflows ready to take advantage of these next generation datasets.
- Moderator
- Will Cadell
- Panelists
- Christopher Lynnes
- David Potere
- Brian Killough
- James Gill
-
Break
Lunch
Pavilion
-
Lightning Talk
Environment
Continental BDeploying deep learning for realtime estimates of hurricane intensity. SAR data for better flood forecasting and flood hazard calculations. Mapping 17 years of forest change in Mato Grosso using machine learning.
-
Lightning Talk
SDGs
Continental CEnding Hunger: Crop yield and condition forecasts for responsive food security.
-
Lightning Talk
Newest in Tech
Hemisphere APlanetary Scale monitoring on a budget with open source and open data. The Cognitive Limitations of AI. Semi-supervised deep learning for rooftop solar potential.
-
Lightning Talk
The Unmapped
Hemisphere BMapping the Missing Millions. High resolution imagery coverage of conflict events. Protecting isolated indigenous communities.
-
Break
Coffee
Pavilion
-
Working Group
Estimating Poverty
Hemisphere AThis session will explore applications of satellite imagery to poverty prediction in developing countries. Participants will share completed work as well as ongoing research across multiple continents. This workshop will compare results from traditional spatial features approaches to newer machine learning approaches. In the final segment, participants will discuss statistical methods and tools for generating local estimates of poverty and average economic well-being by combining survey data, older census data, and indicators derived from satellite imagery.
- Facilitator
- David Locke Newhouse
-
Working Group
The Return of Pixelmonster: Attacking Satellite Image Processing
PavilionPixelmonster (PXM) is a Mapbox service for processing remotely-sensed images and rendering them to maps. This session discusses how to build an imagery processing pipeline at scale in the cloud with API that enables Satellite, Aerial and Drone imagery providers to manage and distribute their imagery at scale.
- Facilitators
- Camilla Mahon
- Todd Smith
-
Break
Coffee
Pavilion
-
Working Group
Responsibly Open
Hemisphere AAn open discussion on working with communities to understand, manage, and mitigate the hazards of open mapping and issues such as mapping uncontacted tribes, consent and privacy, spatial data ownership, and indigenous data sovereignty. How do we understand and quantify harms, including the harm of inaction?
- Facilitators
- Willow Brugh
- Rudo Kemper
- Brian Hettler
-
Working Group
Policy Considerations and Obstacles
Hemisphere BThis interactive session will invite participants to share their experiences related to policy, strategy, and governance opportunities and challenges, with an eye towards informing any needed actions at the national and international policy levels. This is an effort to poll the community on their views of the current legal, regulatory, or organizational barriers to expanding the use of space-derived data for international development.
- Facilitator
- Krystal Wilson
-
Working Group
Open Machine Learning for SDGs
PavilionThe session will propose a collaborative framework for Open Machine Learning approaches for the Sustainable Development Goals. Elements will include creating a global open database of training datasets, standardization and sharing of protocols, and quickly scaling open machine learning tool chains.
- Facilitators
- Keith Garrett
- Caroline Gevaert
-
Break
Coffee
Pavilion
-
Panel
Fast Food Security
Continental CWe have the data to support rapid calibration of crop health and yield, to truly precisely predict food crisis before they start. People on the front lines of food security talk about how we can operationalize food intelligence to prevent hunger.
- Moderator
- Carrie Stokes
- Panelists
- Sarah Muir
- David Johnson
- Steven Brumby
-
Panel
Seeing more than red, green and blue
Hemisphere AHear how these leaders are helping us see through the smoke of wildfires, detect and measure the gas leaking into the air from faulty pipelines, and capture the world at night. This is the tech that we need to go beyond what the capabilities of our human eye and to maximize the capabilities of machines + data from space.
- Moderator
- Charlie Loyd
- Panelists
- Julie Baker
- Yotam Ariel
- Lola Fatoyinbo
-
Panel
Bearing witness to atrocity.
Hemisphere BSatellites are one tool for witnessing atrocities as they unfold, even in the most closed and remote societies. We’ll discuss how satellite analysis, combine with other accounts and data in the quest for justice.
- Moderator
- Mikel Maron
- Panelists
- Julien Cornebise
- Jonathan Drake
- Theresa Harris
- Milena Marin
-
Panel
A pixel tells a thousand words
PavilionUsing precision satellites to get the right shot at the right time and change the lives of thousands of people enslaved at sea. Building the tools for journalists to unite the right images with their investigation to capture the story. Rockstar storytellers share how they’re using maps and imagery to share compelling stories with audiences around the world.
- Moderator
- Al Shaw
- Panelists
- Nancy Coleman
- Christoph Koettl
- Dan Hammer
-
Break
Coffee
Pavilion
-
Panel
Bot or not?
Hemisphere AAI isn’t replacing humans, its helping humans work smarter and faster. Hear from industry leaders in applied AI about emerging design patterns in combining humans and machines to build a queryable planet.
- Moderator
- Chris Knouss
- Panelists
- Tyler Radford
- Mark Wronkiewicz
- Janine Yoong
- Jason Sundram
-
Panel
Conservation Conversation. What’s your relation?
Hemisphere BConservation has led the pack in application of satellite data to their work. People on the cutting edge of this space talk about where the field is going and what other sectors can learn from building realtime global monitoring that informs operations.
- Moderator
- Rachael Petersen
- Panelists
- Jeffrey Allenby
- Jenny Rowland
-
Panel
Out of the Dark. What you Learn from the World at Night
Continental BMost satellites spend half their time over a dark planet. The back pass is often seen as dead time, but its remarkable what you can learn about the world at night. Forest fires, oil production, economic activity.. all shine at night.
- Moderator
- Kwawu Mensan Gaba
- Panelists
- Brian Min
- Kim Baugh
- Frank Riely
- Derek Lieu
-
Fireside chat
GIANTS
PavilionHear great stories from great people who got us to where we are today.
- Moderator
- Bronwyn Agrios
- Panelists
- Darla Werner
- Barbara Ryan
-
Break
Coffee
Pavilion
-
Keynote
Closing
Pavilion
-
Social
After-Hours Party
Smithsonian National Air and Space MuseumJoin us for an evening of off-the-record conversation and an opportunity to hear from former US Astronaut Cady Coleman and former US CTO Megan Smith in an engaging panel discussion of their exciting and impactful careers being badass women working in space and technology. Transportation will be provided. Busses will depart from the 14th street entrance of the Ronald Reagan Building at 6:45pm following the closing session.
ThursdaySep. 20
-
Check-in
Registration & Breakfast
Oceanic Foyer
-
Workshop Track: Analysis Ready Data
Cloud Optimized GeoTIFF
Meridian CThe Cloud-Optimized GeoTIFF (COG) format is a way to store satellite data to allow for fast web access and cloud analysis at scale. Adopted widely by the Earth on AWS effort, it immediately enabled new forms of access and use of imagery data. We will share Mapbox’s experience working with COGs. We’ll also cover how to create them from other formats and how to use them in your application. Difficulty: Introductory/Intermediate. Prerequisites: Python & GDAL knowledge helpful to work on personal data/computer.
- Facilitators
- Vincent Sarago
- Norman Barker
-
Workshop Track: Analysis at Scale
Radiant Earth Foundation platform
Meridian D/ERadiant Earth Foundation will demonstrate their open platform for satellite imagery analysis. Users can access several multi-spectral imagery through the platform including Landsat 8/7/5/4, Sentinel-2, MODIS and ISERV or upload drone and airborne-based imagery to analyze on the platform. Attendees will learn to use the web-based interface as well as the API to work with different imagery and apply analytics to extract intelligence. Machine learning efforts for creating open-source labelled satellite image libraries will be discussed, as well as using Sentinel-1 Synthetic Aperture Radar (SAR) for classification of different features on land. Difficulty: Intermediate. Prerequisites: Basic knowledge of remote sensing & satellite imagery.
- Facilitators
- Hamed Alemohammad
-
Workshop Track: Policy and Strategy
Signal Code, ethical guidelines for humanitarian remote sensing
Oceanic A/BBased on the Signal Code, this workshop will walk participants step-by-step through designing their Code of Conduct and ethical guidelines for a humanitarian or human rights-oriented remote sensing application. Any organization (including businesses) interested in rights-oriented project design should attend. Difficulty: Introductory/Intermediate.
- Facilitators
- Caitlin Howarth
- Isaac Baker
- Saira Khan
-
Break
Coffee
Oceanic Foyer
-
Workshop Track: Analysis Ready Data
DataCubes and CommonSensing
Meridian CParticipants will learn about how to access and exploit EO data stored in an Open Data Cube (ODC) to generate data related to the Sustainable Development Agenda. Data Cubes are a new way to store, organise, manage and analyse EO data. Data Cubes provide access to large spatio-temporal archives that are processed and organized to allow immediate analysis across time and across data sets with minimum user effort. This removes barriers to full exploitation of EO data and supports use by non-expert users. For this hands-on workshop we will discuss the ODC and it’s role in facilitating improved climate and disaster risk resilience in Small Island Developing States. The Radiant Earth Foundation will introduce the Common Sensing Project, and outline our vision for an open source solution that will be deployed in the SIDS. NASA CEOS will demonstrate the ODC, and attendees will have the chance to access a ‘Sentinel-1 Data Cube Playground’ for Samoa (Please bring your laptop!). Difficulty: Intermediate. Prerequisites: Basic knowledge of satellite data. Knowledge about radar data and Python coding will yield advanced value.
- Facilitators
- Terri Freemantle
- Asimina Syriou
- Brian Killough
- Anthony Burn
-
Workshop Track: Analysis at Scale
Rasterio
Meridian D/ELearn how installation of the satellite data processing stack is used at Mapbox, how to use the Rasterio library, and how to integrate Rasterio with Numpy and cloud platform libraries to access and transform satellite data. Difficulty: Intermediate. Prerequisites: Cloud computing & Python.
- Facilitator
- Jacques Tardie
-
Workshop Track: Machine Learning
GBDX
Oceanic A/BLearn how DigitalGlobe is using Jupyter Notebooks and GDBX to create an open data science community for sharing models and structured output, and how to leverage the open source ecosystems like Jupyter to develop custom communities. Explore challenges and solutions of integrating large data repositories with an open analytics stack and creative approaches to on-boading new users to satellite imagery and machine learning for collaborative work. Difficulty: Intermediate. Prerequisites: Basic understanding of remote sensing & ability to program in a modern programming language (Java, JavaScript, Python…)
- Facilitator
- Steve Pousty
-
Break
Coffee
Oceanic Foyer
-
Workshop Track: Analysis Ready Data
Google Earth Engine and Python
Meridian CIn this workshop, we will walk attendees through obtaining satellite imagery, processing and preparing it for analysis, and linking it to analysis engines and platforms. Participants will prepare ARD data using two repeatable methodologies: one cloud-native workflow using Google Earth Engine; and the other using open-source Python tools and libraries on a local environment. For this hands-on workshop participants will access open imagery from Planet Labs, Landsat and Sentinel-2. Difficulty: Introductory/Intermediate. Prerequisites: Familiarity with python is a plus, but not necessary. Please bring a laptop to access materials and follow along.
- Facilitator
- Samapriya Roy
-
Workshop Track: Analysis at Scale
sat-utils
Meridian D/EThe session aims to demonstrate to participants how to use the sat-utils collection of tools to discover and access publicly available imagery such as Landsat-8 and Sentinel-2. The latest versions of sat-utils take advantage of the new SpatioTemporal Asset Catalog (STAC) spec, which enables a standardized workflow for discovery and processing of data, regardless of the sensor it is from. Several sat-utils programs will be explored over the course of the workshop. Sat-api is an easily deployable infrastructure to run your own STAC compliant API. We run a public version of sat-api for all the Landsat-8 and Sentinel-2 data on AWS. Sat-search is a library and command line tool for searching and downloading data from sat-api, while sat-fetch allows for downloading just your area of interest. During this workshop you will learn how to use these tools to quickly find Landat-8 and Sentinel-2 imagery for any AOI, and fetch the data for processing.
- Facilitator
- Matthew Hanson
-
Workshop Track: Machine Learning
Label Maker and Robosat
Oceanic A/BThere are a handful of existing annotated datasets for satellite imagery, but they’re mostly limited to urban regions in Western countries. This presents a challenge when attempting to use machine learning algorithms in non-Western regions. This workshop will show you how to generate custom data for feature extraction from aerial and satellite imagery. We’ll use two open source tools, Label Maker and RobotSat, to prepare the data, train an algorithm, and predict over a test area. Difficulty: Intermediate. Prerequisites: Basic understanding of machine learning, satellite imagery, and OpenStreetMap.
- Facilitators
- Drew Bollinger
- Jacques Tardie
- Mark Wronkiewicz
-
Mapathon
OpenStreetMap Hackathon for HOT and Detroit
Mapbox DCCome grab a hot dinner and drinks with other map enthusiasts after SatSummit on Thursday Sept 20, 6pm-8pm, for an OpenStreetMap mapathon where we will help map Detroit as part of the Detroit Mapping Challenge and urgent humanitarian mapping needs as part of Humanitarian OpenStreetMap. No prior OpenStreetMap editing experience required - we’ll teach you everything you need to know. Just bring your laptop.