Once again, it’s time for another one of my conference “mega posts,” this time from the 2016 UCLA Mobile & Modern Web Conference. This mega post links to the notes I took from almost every session I attended, and proves that I was here and learned something. Some of this will be more useful than others…I drifted at times. There was some great work on display at this conference; I always meet awesome people and take away at least a couple gems that I didn’t know going in. As always, any mistakes, omissions or just plain crappy coverage are totally mine. Who knows, I may have even gotten something right 😉 Enjoy!
Wednesday, October 14
Thursday, October 15
Friday, October 16
- Curtis Bray, Manager of Solutions Architecture – US Education
- Heather Matheson, Account Manager for UCLA
- Amie Carobrese, Account Manager for UC & CSU
We’re deep into the public sector! 2,300+ govt agencies, 7,000+ educational institutions, etc.
When starting out, you pick a region, which is a location where your stuff (apps, data, whatever) will live. For example US-West-2 is a region. Within each region is a number of availability zones (AZs), separated by tens of miles.
How Do You Build a Mobile App Today?
- Store & Deliver Media
- Analyze User Behavior
- Send Push Notifications (bus arrival, quiz/test results, etc.)
- Send Real Time Events
- Store Shared Data (i.e. leaderboards)
Some of the products that’ll be covered:
- Cognito: authentication
- Kinesis Recorder: create live dashboards to track user events in real time
- DynamoDB Mapper: store & query fast NoSQL data across users & devices
- SNS Mobile Push: for push notifications
- Lambda: “serverless” service that runs your functions
Authenticate Users: Amazon Cognito
What it does:
- Synchronize user’s data across devices / platforms
- Manage users as unique identities across identity providers
- Securely access all AWS services from Mobile device
It can provide temporary credentials to securely access your resources. It provides comprehensive support for identity use cases.
Synchronize data across devices: Amazon Cognito (Sync)
- Store app data, preferences and staet
- Cross-device Cross-OS Sync
- Work Offline
Post processing can be done: push sync, events, streams
Store & Deliver Media Assets: S3 & CloudFront
- S3 Connector transfer utility: multipart upload, fault tolerant download, no backend required, automatic retries, pause, resume, cancel functions.
Collect, visualize and understand app usage data. Data stored can be put into a data warehouse if you want to.
Send Push Notifications: Amazon SNS Mobile Push
A single API can push messaging across multiple platforms.
Amazon DynamoDB Connector: Object Mapper
A NoSQL database that’s mobile aware
Which services should I use?
How do I connect them all together? We created the “AWS Mobile Hub” which ties into your AWS account. You create a project and add features you need to get up-and-running. The AWS Mobile Hub needs an IAM role so it can create resources on your behalf.
Working Mobile App project includes:
- Xcode/Android Studio project
- AWS resources
- App Code
- Helper “glue” code
- Detailed developer instructions
We then ran through an hour-long demonstration of the tools mentioned above…pretty powerful and easy-to-use stuff. The automated testing with AWS device farm was cool. You test against actual physical devices.
Presenter: Alfonso Roman
I’ve played with a lot of the watches.
- They don’t work on their own
- Companion apps handle communication with your smart phone
- Cloud based account to use core functionality (i.e. messaging)
UX & UI
- Each has it’s own design & interaction language
- Rely heavily on touch, gestures, and voice
- Different interaction affordances
- Android Wear defines notion of a context stream that uses vertical swipes to view items vying for your attention
- Each requires their own SDK
- SDKs provide hooks into OS features (Apple Health, Google Fit)
- Pebble SDK provides limited bridge to iOS and Android apps
- Fitbit SDK is a RESTful API; not possible to talk directly to the Fitbit device, all communication must happen through the cloud after the Fitbit has synced data with its servers. Can sometimes get stale data based on sync schedule.
- Battery life (Pebble & Fitbit are great)
- Native vs Web API
- Dev platform
- Integration with other frameworks
- Apple watch implies native iOS app
- Android watch implies Android Phone
- FitBit is web only
- Pebble: just plain ugly (C++ only)
- Companion apps: devices are useless without them
Presenter: Arthur Lobzinki, Oomnitza
- A “Thing Management Organization”
- We have about 20 people globally
- Manage/track lots of stuff (mobile devices, servers, self-driving cars, etc.)
- Lots of big-name companies
Internet of Things
- Before the industrial revolution, it was typical for most people to have only two shirts, which seems ridiculous to us now.
- Before the information revolution, it was typical for large clunky computers to be the
- Computers are now built into everything
- Cisco will spend $1.3 TRILLION on the IoT
- SIM cards are going into everything – spits out data about the devices they’re installed in
- GE wind turbines have 17 servers built into them!
- LinkNYC: 400 public internet machines have been built which provide GB service, charging station, and advertising space
- Uber is largest puchaser of 2nd hand iPhones; they have 20 self-driving cars in Pittsburgh.
- Agriculture: supply chains, crop irrigation patterns, etc.
- Mining: equipment/asset uptime. Just keeping things running is the number one most important thing for this industry.
- Security! If everything is connected and and endpoint, it’s vulnerable.
- 97% of all hacks are human error :-/ and directly connected to enterprise systems.
- Employment: a lot of jobs will be replaced. What do you do with all the people? Fixing the machines.
- The only real value in IoT that we’ve seen in the industrial realm is in keeping the machines up and running.
Lots to think about here!
Presenter: David Ahn, UCLA Endocrinology @AnhCall
My clinical career and the iPhone were very intertwined, and I’ve been very interested in the use of mobile technology in medicine.
Medicine versus Business
- They’re very different; polar opposites in many ways (both have their pros and cons).
- Business-led ideas in medical hardware are often not what they’re promised to be (over promise & under deliver), i.e. Scanadu, Theranos: Hubris in disruption.
- Medicine-led ideas: clinical utility over the user experience (“everything but the sink” approach), no clear business model: hubris in being a domain expert.
Achilles Heel of Digital Health: Engagement
- Poor user engagement
- Preaching to the choir
- What happens after 30 days?
- Do they actually lead to behavior change
- Passive measurement alone is not the solution
How Do We Engage & Motivate Users?
- Habit-forming psychology
- Variable rewards, i.e. slot machines
- Loss of reward > Hope of earning reward; expiring offers
Apps that Do a Good Job with Engagement
- GymPact: “get paid to work out,” commitment-based, negative & positive reinforcement, effort-dependent, not result-dependent
- Prevent by Omada Health: 16 week online course structured around the DPP (I think this one is based on solid medial principles); health coach via phone, text, private message; social motivation (teamed with 12-18 random people); technology tools provided include a wireless scale and pedometer.
Hipster / Hacker / Hustler
- UX, Engineer/dev, marketing/business
- In health startups you also need the healer!
- 29.1 million in US have it
- Hundreds of millions are spent on it
- Managing it: monitoring glucose, diet, physical activity, exercise
- “What gets measured gets managed”
- Regularly checking sugar is proven to help improve diabetes control
- Traditional: use glucose logbooks (a vital tool)
- Every device has it’s own proprietary plug/cable
- Information overload
- I don’t want to feel bad
- I don’t gain anything from it
- Insurance restrictions
How to Change Perspective?
- Scare tactics
- Social pressure
What Works for Me
- “Pact App for Diabetes?
- Commitment based
- Negative and positive reinforcement
- Effort-dependent, not result-dependent
This approach helps to build a healthy habit
- Build streaks
- Learn and earn
Trends and habits are reflected via graphs
App has been successful so far!
v1.8 New for “Code for the Mission”
- Dashboard for viewing/sharing glucose reports
- Viewable on an y mobile or desktop browser
- Notes field
How are we doing on User Engagement?
- 10% log in every day
- Longest user streak: 164 days
- Thousands of users have logged 50k sugar readings
David then gave a demo of the app