2015年6月28日日曜日

SF Hardware Startups Meetup

I joined SF Hardware Startups Meetup this week, here are some pics.

Format of the event is that each startup gets 2 minutes, no slides are allowed (in fact they won't have projectors) so you either demo or talk.

The fun part is before and after the talks, you have tons of time to touch their products in the early stage (usually not on sale yet), try them, play with them, and talk with the teams that built them :)

Bolt motorbikes and their electric motorbike.

SF Hardware Startup Meetup

It has several modes: On "Economy Mode", it is 20 MPH top speed, 1000 watts of power and you can drive up to 50 miles range. On "Sport Mode" it's 40 MPH top speed, up to 5500 watts of power and can ride for 30 mile range. It has regenerative braking too.

SF Hardware Startup Meetup

SF Hardware Startup Meetup

It connects with your smartphone app via bluetooth, you can charge your smartphone as well.

SF Hardware Startup Meetup

The seat looks hard on picture but when you sit on it the cushion is very soft.

SF Hardware Startup Meetup

SF Hardware Startup Meetup

You can make a reservation or request a test ride from their website.

PureRockets showcased their TurbineDrone. It has 4 turbines, the silver rod you see in the picture connecting the turbine and center- that's the fuel. It can fly as fast as 350MPH.

SF Hardware Startup Meetup

SF Hardware Startup Meetup

SF Hardware Startup Meetup

This is how you'd look when you fly. He'll be flying in 3 months, and is planning to launch Indiegogo campaign in 6 months to allow others to fly.

SF Hardware Startup Meetup

SLICK is a stabilizer for GoPro.

SF Hardware Startup Meetup

Video of GoPro footage with and without SLICK.



They're planning to start Indiegogo campaign soon.

Sensel is a pressure sensor tablet. Based on the level of pressure you put on the tablet, you can visualize that force.



SF Hardware Startup Meetup

SF Hardware Startup Meetup

They're working on getting their first units to production, so stay tuned!

Titanium Falcon team is building a motion control ring that works on iOS or Android. (This is just a mock)

SF Hardware Startup Meetup

Awair analyzes indoor air quality in real time, communicate with other home devices to help you achieve optimal air quality.

SF Hardware Startup Meetup

Others who spoke includes Indiegogo's hardware team, OpenbikeStratio (imaging system), Whatweorder (business purchasing cloud software),  discoverGPiO (hardware component search), Anova Culinary (precision cooker), portable box to get rid of wrinkles, search engine for hardware, hardware startup accelerator, consultant, etc.

Thank you to Cisco Meraki,  the host and sponsor of the event this month, and they gave us Z1 as gifts- yay :D

SF Hardware Startup Meetup

Meraki had a beautiful office, their patio has a great view- and they're hiring!

SF Hardware Startup Meetup

Thanks to the organizers- it was an awesome event!
Other photos I took from this meetup can be found here.

Disclaimer: The opinions expressed here are my own, and do not reflect those of my employer. -Fumi Yamazaki

2015年6月23日火曜日

Smart Cities, Smart Communities (Notes from Technologies for Crisis Response & Community Resilience #4)

Sensors in Service of Public Health 
Evan Thomas (Portland State University)

Examples of Interventions in Rwanda

IMG_20150622_135105

Example1. Sensor of water filter use, looking at impact on diarrhea

- By using sensors, we can actually know the use and behaviors, which is different from surveys-  there are huge discrepancy between actual sensor data and survey result

IMG_20150622_135318

- Seeing inside the households: How consistently are they using water filter?
  ex. replied they are using filters but reality is only once in 2 weeks
- Water filter use between open and closed arms > when people knows there are sensors, they behave differently

Example 2. Pump in Rwanda

- In the past, staffs were looking for broken pumps vs using sensors and based on data sending people to fix pumps
- Cost effective? yes

IMG_20150622_135656

IMG_20150622_135909

- This picture does not mean those children has clean water!

IMG_20150622_135944

Building Resilient Health Systems Through Remote Monitoring and Analytics 
Martin Lukac (Nexleaf Analytics)

- 20 million children vaccinated
- Vaccines are too hot or too cold, because fridge is malfunctioning
- "Why not replace fridges in poorest countries?"
- No good data about fridges' performance
- Sensors to send out alerts when temperature goes wrong, dashboard
- Testing with few hundred fridges, on average 50% of the time, too hot or too cold
- New fridges are failing as much as older ones > replacing old fridges are not going to solve the problem
- Implemented SMS alert > improved 10x
- Vaccine fridges can call for help
- Data-driven procurement decisions
- Resource allocation: fridge, parts, technicians, workflows
- Greater vaccine coverage at same costs, more kids vaccinated
- IoT for developing countries -> focus on communities, groundwork for resilient communities
- Instead of taking things built for developed countries and bringing it to developing countries, build with developing countries

IMG_20150622_141430

IMG_20150622_141755

Bridging the Environmental Data Divide 
Sean McDonald (Frontline SMS)

- Challenge - need fixed line power or Internet signal
- DustDuino -> OpenDustMap: measuring air quality by public transportation, testing in Brazil

IMG_20150622_142635

- Resilience is an organizing challenge -> data is not output, it defines the process, need meaningful interaction
- Every decision about data is a decision about power
- Communities are the hard part
- Organizing communities begins by meeting people
- Less than 40% of people in the world are not online
- 7.2 billion active mobile subscribers, 3.6 billion unique users - SMS is the only way to reach them all
- Digital tech makes them better informed than 5 years ago
- People in the country with Ebola -> getting 200 messages saying "wash your hands" :(
- Context is important
- Resilience is not one message away! Relationships building is the work.
- No extra time in disaster- embed context and significance into routing and processing of the data in the system.
- We cannot assume homogeneity. Resilience approaches must be multi-channel.

Message powered healthcare: Hospital in Malau doubled the number of patients they can treat

IMG_20150622_143601

Message powered education

IMG_20150622_143630

Bolivia Mercy Corps, Networked land titling

IMG_20150622_143653

Interactive Radio
- Radio reaches the most number of people during disasters

IMG_20150622_143726

Message-powered banking

IMG_20150622_143746

Disclaimer: The opinions expressed here are my own, and do not reflect those of my employer. -Fumi Yamazaki

Ignite Talks: Perspectives from Implementers (Notes from Technologies for Crisis Response & Community Resilience #3)

Early-warning Systems to Prevent Fires in Urban Slums 
Paul Mesarcik (Lumkani)

- Fire spreads quickly, houses are fragile to fire and extremely dense in those urban slums
- In March 2015 in Manila, just 2 fires made 10,000 families go homeless

IMG_20150622_122031

Solution:
1. In-home heat detector
2. Transmission to neighbouring devices
3. Data collection through internet gateway

2007 - 1 billion people worldwide were affected
2020 - 1.4 billion people worldwide will be affected

Urban slums are one of the most underserved communities in the world.

Peer to Peer Communication for Resilience 
Meena Palaniappan (Atma Connect)

- Connecting and empowering people for resilience
- People are not objects of development, we need to work together
- Poor people will suffer the most from climate changes- flooding, etc
- Social cohesion is key resilience tool (people are far more likely to survive disasters when they have good social networks and connections)
- Mobiles for people- build technology to help people solve their own problems.
- AtmaGo - a mobile web app to connect and help neighbors
- People in poor neighborhoods miss early warnings. Hyper local information sharing is important, with information such as flood is coming, road blocks, etc.
- AtmaGo is built on Facebook and Twitter, since that's what people in Jakarta use.

IMG_20150622_123012

Data Mining to Increase Small Farmer Resilience & Agricultural Performance in Latin America 
Everett Wetchler (Bayes Impact)

- Agriculture yields problem- focus on rice
- Growing global competition, climate change, FTA making farmers difficult
- Organizations in LatAm- many are funded internationally
- Big data in rice agriculture workshop - many talented agronomists, very few software engineers

IMG_20150622_123520

Finding Data Invisibles: Addressing the Data Gap with Callboxes 
Rose Shuman & Mita Paramita (QuestionBox)

- BrightFront Group
- National Ebola Hotline for UNICEF
- QuestionBox infrastructure is free public community for central hotline in local language
- Featured in World Economic Forum as data-driven development
- Community based reciprocity
- Putting QuestionBox in health clinics

IMG_20150622_124216

3 Modes of resilience
1) direct linkages 2) self-sufficiency 3) short term infrastructure

Crowdsourcing Community Resilience Solutions in Dhaka 
Ian O’Donnell (Red Cross)

- Safer, healthier and stronger communities
- Designing across Red Cross Red Crescent Network: 189 countries, 60k local branches, 17 million volunteers
- Engaging one billion people for safety and resilience by 2025
- Enabling local problem solving at scale
   -RedCross, Governmental networks, academic and NGOs, business networks
- Making city coalitions
- Working with urban slum in Dhaka. Not a lot of information sharing
- Working with Open Streetmap, Tech Hubs, utilization of mobile data, working with World Bank, etc

IMG_20150622_125359

Disclaimer: The opinions expressed here are my own, and do not reflect those of my employer. -Fumi Yamazaki

Collaborative Mapping & Crowdsourcing (Notes from Technologies for Crisis Response & Community Resilience #2)

Integrating Mobile Data Collection with Satellite Imagery 
Nicholai Lidow (Premise)

- Remote sensing and mobile data collection
- Detect many types of economic activity
- Less than 25km between Siera Lione and Liberia near the border, but the patterns of crops are changing

  IMG_20150622_114412 

- No matter how good the sensors get, there are things you cannot see fro the satellite (ex. indoors, cloudy, higher frequency, greater resolution) 

Example: India drought 

- Variation in cold storage infrastructure and strategic grain reserves had bigger impact on prices than drought itself
- Collecting rapid information as needed

IMG_20150622_114627 

Example: Indonesia

Global coverage of mobile data
- Currently in 30 countries, expanding to 140 countries in 2 years.
- Geotagged photos, getting the data from contributors and partners
- Data update by untrained contributors- will judge over time, teams looking with human insights to potential data quality issues.
- Recruiting contributors via job boards, SNS, partner with local organizations, sending people in if needed.

IMG_20150622_115003

- Creating a feedback loop- global network of mobile data collection and standardized data stream, analytics and providing additional data to improve forecasts and answer questions.

Using Telecom Data for Social Good 
Daan Struyven (Real Impact Analytics)

- Using big data to help policy makers and aid workers foster the social public good and maximize impact

1. Improving the response to disease outbreaks

- Pandemics killed, and can kill millions of people
- Bill Gates pushing research
- Lessons from Ebola
  - monitoring could have been better
  - medical and supply
  - quarantines (quarantine rules were completely ad hoc)
- Developing 3 apps: disease monitoring app, medical supply app and quarantine app

IMG_20150622_120446

2. Detect and tackle food crises through better monitoring

- 800 million people are under-nourished
- Cause is food demand by low and volatile income / supply (volatile climate, crop quality)
- Income shocks and nutrition ($2/day)
- Food security monitoring

IMG_20150622_120739

3. Building broader and open ecosystem to co-design

IMG_20150622_120802

IMG_20150622_120914

Q: How do you get data from mobile companies?
A: Compensating for cost, changing culture of data sharing to use it for public goods

Disclaimer: The opinions expressed here are my own, and do not reflect those of my employer. -Fumi Yamazaki

Social & Environmental Monitoring from Space (Notes from Technologies for Crisis Response & Community Resilience #1)

Today I joined Technologies for Crisis Response  & Community Resilience meeting at Berkeley and here are some notes.

Agenda of the conference:

Democratizing Access to Information About our Changing Planet 
Andrew Zolli (Planet Labs)

Our collective resilience challenge:
- How do we help communities and systems persist, recover and thrive amid disruption?
- How do we make communities listen? > sociological issue, not technical issue.
- We should be in age of big indicator, NOT age of big data.

Successful crisis response:
- Context specific
- Bound together in ad hoc manner (like Jazz)
- No single enterprise, tool, or organization can solve everything in disasters and resilience.

Crisis response should be considered in broader time horizon
- Natural or manmade disaster, refugees used to be 8 years in past, increased to 20 years (baby gets adult in shelters...)

Planet Labs > aiming to get visuals everywhere, every day, will enable a new approach to water, nature, city planning.

We can make API today, but everything might change tomorrow. It is more important to do capacity building. We need to work "with not for" the communities.

NGOs- need a new model of funding, collaborating, experimenting and sharing.

Newest things are always the most fragile. Old technology is not sophisticated but robust. Make the new technologies less fragile by working with the existing systems. Not about Silicon Valley way of disruption, we need to be working with existing processes. We need to be agile enough to work in different ways and adjust to local communities.

IMG_20150622_094705

Assessing Impacts in Agriculture at Ultra-low Costs
LSMS (Living Standards Measurement Study)
David Lobell (Stanford University) & Talip Kilic (World Bank)

-3 elements for ultra-low cost accurate crop monitoring
1) Satellite data 2) Computation (Google Earth Engine) 3) Robust algorithms

Joined Skybox for Good program

Testing in 4 areas: West Kenya, East Uganda, Rwanda and NE India.

IMG_20150622_100619

Example in Kenya
- Got imagery and walked the fields
- Vegetation, measured the facts and on the ground information and data

Example in India
- Crop cuts

Example in Uganda
- Development Research Group at World Bank

Maize Productivity, variety and soil fertility (MAPS)
- Supported by LSMS methodological validation program
- Test subjective approaches to measurement vis a vis objective methods for maize production, variety and soil fertility
- Crop cutting is gold standard
- Physical measurement

IMG_20150622_100824

MAPS methodologies

IMG_20150622_101351

- GPS data vs data that farmers are reporting (farmers are over estimating)

IMG_20150622_101638

IMG_20150622_101745

IMG_20150622_101759

Difficulty in data: Even in this quadrant, crop density is very different

IMG_20150622_101930

- In the past: lack of attention to data quality
- High frequency surveys of households and communities at strategically selected locations

Differentiating Irrigated and Rainfed Agriculture for Efficient Water Usage in Morocco 
Travis Lybbert & Michael Norton (UC Davis)

Climate change: notice change and then adapt... resilience requires recognition that "something changed". How well and quickly can farmers detect those changes? A lot of noise, complicated learning, tough learning process.

Looking at Kilimanjaro everyday- the farmers has the history. But there's good years and bad years, how can farmers detect the bigger changes? Farmer-level decision making. Can we track the spatial and temporal diffusion of irrigated agriculture in recent decades using satellite images?

IMG_20150622_103532

Existing data and methods:
-Moroccan Agricultural Atlas and FAO AQUASTAT raster files, etc
-Methods: land cover change detection (comparison of 2-3 scenes to using them all)
-Continuous Change Detection and Classification (CCDC) Algorithm

Seasonal patterns

IMG_20150622_104149

IMG_20150622_104208

IMG_20150622_104336

IMG_20150622_104410

High-resolution Earth Imagery for Sustainable Development 
Taner Kodanaz (DigitalGlobe)

- Better understand and better seeing, solving the problems, not just consuming imagery but working to bring them into existing systems.
- Small and agile organizations do better at utilizing those data, but when come into national, UN, World Bank level, things are more difficult and we need to find out how to collaborate.

Global challenges
Agriculture, environment, human rights, infrastructure development

IMG_20150622_105650

Digital Globe is focused on how to help organizations make better decisions by giving the customers the power to see the earth clearly and in new ways, enable them to make a better place whether it is through increasing efficiency or saving lives, etc.

What satellites can see and can't see:

Panchromatic imagery (black and white) -trees, minerals, etc
4-band imagery - differentiate evergreen and deciduous tree
8-band imagery - pine, cedar etc

IMG_20150622_110049

SWIR imagery (short wave infrared)

IMG_20150622_110127

UN's 17 sustainable development goals:

IMG_20150622_110232

Through those imagery, we can see individuals, where people are living and moving, and the infrastructure to meet those goals.

IMG_20150622_110329

IMG_20150622_110414

"The surface of Earth itself is an immense loom upon which the sun weaves the fabric of existence"

Q: How do we include the communities and right insights that they have? They are the ones who has insights and implementation for resilience, not the ones who are working on data and algorithms

A: Working together with communities

Disclaimer: The opinions expressed here are my own, and do not reflect those of my employer. -Fumi Yamazaki