technical disruption

Personal Blog of Michael Jones aka MJ Maker + Digital + Physical + Bigdata + Analytics + Social + Ruby + Arduino + Adventurer Follow @mjfreshyfresh !function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0];if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src="//platform.twitter.com/widgets.js";fjs.parentNode.insertBefore(js,fjs);}}(document,"script","twitter-wjs");

[mjfreshyfresh]

If you own an imported grey market vehcile it's quite likely it's in kilometers per hour and you're always wondering "What do I need for doing 25 mph in this speed zone?". Well here you go, print this out and tape it to your windshield. Perhaps someday I'll make it a sticker.

[mjfreshyfresh]

Connected Devices: Hierarchy of Capabilities

Having researched the current state of IoT / Connected device / Smart Device / Connected Products / Etc... which they are predicting 75 billion devices by 2020 I've come to a few conclusions.

Hardware has matured greatly, a lot of the "hard" problems have been solved. Hardware will become a commodity.

IoT will ultimately be about the data and software.

Despite rapid advances in hardware and software we are just starting to comprehend the "tip of the iceberg".

Ethics and security will be infinitely more complicated in this new "connected world".

The future is going to be exciting! I put together a little pyramid diagram to help frame up connected capabilities. Next time you come across an activity tracker, smart yoga mat, learning thermostat or other connected product think a bit more about what is actually happening.

Is this device just monitoring something? Does it use the data to control something else? Is it smart enough to optimize for a particular outcome? Is it completely autonomous?

IoT makes these capabilities a reality only limited by our imaginagtions. Move up the pyramid to unlock greater value. You must learn to walk before you run. But with the amazing innovations in hardware it's much easier to get up and running.

[mjfreshyfresh]

Data Analysis to Catch Fraud

An ideal Saturday morning for me includes sipping a cup of Keemun tea and having plenty of time to read the paper. This morning was an ideal Saturday and I was really intrigued with the Accountants Increasingly Use Data Analysis to Catch Fraud article by JO CRAVEN MCGINTY

I was surprised to find that employee fraud is estimated to be around $300 billion a year. I completely agree with Timothy Hedley's quote "The future of forensic accounting lies in data analytics,". But the item that really got me excited was Benford's Law aka "First-Digit Law" which refers to the frequency distribution of digits in many real-life data sets. For instance 1 occurs as the first digit about 30% of the time, from there each digit appears less frequently. It goes like this:

1: 30.1%

2: 17.6%

3: 12.5%

4: 9.7%

5: 7.9%

6: 6.7%

7: 5.8%

8: 5.1%

9: 4.6%

Given my love of data and analytics, I quickly set off to test Benford's Law on with my own financial data! Mint made downloading the past 4 years of financial transactions as simple as a single click. From there I used a bit of Ruby to manipulate the data and finally I used CHARTED to visualize my "actual" data vs Benford's "expected".

Most digits tracked closely to the expected value. However some numbers for example 5 appeared more frequently than expected so I set about trying to figure out why. I parsed out all transactions starting with a 5 and reviewed things. Turns out a lot of my automatic payments and investments start with a 5 which over time could over skew the data.

I thought about wrapping this whole process up as a web-app so others could check their transactions but in the end I didn't think folks would want to paste financial records in a random web-app. So I'll leave you with this live chart feel free to contact me if you're interested in running your own numbers.

[mjfreshyfresh]

Building a power supply for electronic projects

"Electricity is really just organized lightning.” - George Carlin

As you start hacking you’ll quickly realize "organized lighting" is important and comes in many shades. My bike sound system and PiLights take 12 volts. Where as Adafruit NeoPixels takes 5 volts and some things even take 3.3 volts.

If you’ve got an old ATX computer sitting around your house then you’ve got what you need to build a power supply like the one I built pictured above.

There is a ton of detailed information on the web, personally I found this Instructables post quite helpful. Few things I might add:

You can easily buy a power supply ~$10 used.

Most supplies that have colored wires, stick with those instead of vanity ones where all wires are the same color. (silver)

Remember you’ll need to loop the green wired back to ground.

Remember you’ll need to loop the brown wired back to 3.3v.

If the power supply box is small then buy a "project box” ~$3 at Radioshack to mount the plugs.

[mjfreshyfresh]

Iot: More Value Less Novelty

I love “novelty”. Unconventional, unfamiliar, entertaining bring it on! Some technology appears to be a novelty however, over time application of it can mature into something that creates value. The Internet of Things / Internet of Connected Products whatever you call it, this is something I feel is on track to move from novelty to creating great value.

Which brings me to @PiLights. This Summer I gave people the ability to tweet commands at a 6 large lights hung in a big tree. People were utterly amazed! They laughed, they squealed, they tweeted! I’ve even got neighbors who had never Tweeted to do Twitter. Simply because I created something that was unconventional, unfamiliar and down right entertaining. A NOVELTY.

Sadly at this moment PiLights merely perpetuates this perception that the #Iot is a “novelty” and on the contrary I want to demonstrate how it came make life better. This is how I intent to move from novelty to value...

Lighting - Provides ample light for entering/exiting house. Never worry about having to leave a light on or wasting electricity. (Knows date/time and sunset/sunrise. Tap weather API for cloudy days.)

Security - Collaborate with my Dropcam depending on time and movement start flashing mode to confound.

Information - Provide fade based rotating visualization of current tide and next 3 days. (Good to know for crabbing and clamming.)

Fun - Twitter “friends” of @PiLights can still tweet to unlock new light activities.

[mjfreshyfresh]

Parenting: Graph for Success

As a step-father I've got a unique relationship with my kids. (23 and 19 yrs old) I'm more of a mentor providing advice on tough decisions, cheering them on towards success and at times having brutally honest conversations. The other day my step-daugther and I were discussing some of her recent actions. I was helping her think about how she could make better decisions, with an eye towards increase opportunities first and money second.

Given my love of data visualization I crafted a simple graph to provide her with a tool to assist in decision making. She found this quite useful so I decided to share it incase there were other parents out there looking for ways to arm their children for success.

X-axis: Cost - Is this action making money or costing money? (Note: Costing money is not necessarily bad)

Y-axis: Growth - Is this action growing your opportunities, skills and knowledge?

Based on these two axes you get four quadrants. Ideally stay away from the bottom-left and move towards the top-right. However the realistic approach is to start from either the top-left or bottom-right and over time move towards the top-right.

Top-right: Makes money and creates opportunity. This is the ideal but typically not where you start. Examples: A relevant job or paid internship.

Top-left: Costs money and creates opportunity. Most likely where you start. Examples: Meet-ups, Study groups, school.

Bottom-right: Makes money and doesn't create a lot of opportunity growth. Another place where you can start. Examples: bar tending, drug dealing

Bottom-left: Costs money and doesn't create a lot of opportunity growth, on the contrary it destroys opportunity. Avoid! Examples: Hard partying, jail.

[mjfreshyfresh]

IoT Architecture Patterns - Sensors and Actuators

I came across Arduino around 2008. After spending many years of my life writing code that did little more than manipulate pixels on the screen the ability to make things happen in REAL LIFE was intoxicating!

It started with a tweeting pinewood derby car then I got into shocking people at World Maker Faire and lately I've been tinkering on an interactive led project.

All of these projects fall into the Internet of Things category which is essentially connecting devices. By connecting devices we open up an entirely new world of insights and activities! We're essentially sitting on the cusp of the second industrial revolution!

Earlier this month I was at Dreamforce and was astonished at the resources Salesforce has put towards IoT or as they like to call it IoCP (Internet of Connected Products). Using Little Bird I quickly found Reid Carlberg and have been enjoying his content ever since!

Just the other day Reid published Five Enterprise Internet of Things IoT System Architecture Patterns and given my love of dataviz I quickly put them into a scatter plot. From this diagram it's clear that as report frequency increases, Buffered and Gateway patterns become obvious solutions.

I have nowhere near the level of expertise as Reid but after reading his article I started to think about different ways to classify IoT systems. His nervous system metaphor really stuck with me so I sketched out a way to classify IoT needs based upon:

Sensors (reporting) - Frequency of (high/low) and payload of (large/small)

Actuators (action) - Type of (human/machine) and complexity (high/low)

[mjfreshyfresh]

Overland Influencers that are Under-the-radar

'Life is a journey, not a destination.' This is my favorite quote from Ralph Waldo Emerson and it's the essence of "Overlanding". My love of adventure, travel and engineering is what draws me to Overlanding. Have a look at my Overland Exploration collection on Medium and you'll see what I mean.

I'm also into data and technology. As CTO of Little Bird I spend my days analyzing social networks and developing approaches to help extract valuable information from the social graph.

The other week I decided to use this social graph approach to identify the best accounts to follow on Twitter for Overland information. From that I created a Top 10 Overland Adventure List and finally a TreeMap. Several people asked me how I identified these accounts, so here it goes...

Most companies these days use primitive measures to determine "influence". They typically look at 1. Who has the most followers? 2. Who says particular keyword? with this approach you end up with folks like Kim Kardashian and Justin Bieber being influencers in practically everything.

For my top 10 list I started with a Little Bird report on Overlanding. The way this report works is it starts with some "seeds" which are examples I provided of exemplary Overland accounts. From there the software crawls the social graph looking at RELATIONSHIPS not follower counts or keywords. (friends, friend-of-friends, etc). At that point we calculate indegrees which simply put is "contextual popularity" not "absolute popularity". Of all these Overland accounts which accounts garner the most respect from other Overland accounts? It's a simple form of peer validation.

To get to my top 10 I didn't just stop when I had the accounts with the highest indegrees. (If I had stopped then my list would have been folks like @earthXplorer @lonelyplanet @NatGeoTravel @vagabondish @earthXplorer ). What I was looking for was the "under the radar" type of accounts. Those accounts that may not have hit "main stream" just yet but have garnered a ton of respect from the community. To do this I took the total indegrees and divided it by the total followers. (contextual popularity / absolute popularity) and there I found what I was after:@the4x4podcast @Motojournalism @worldcyclevideo @OverlandExperts @AmericanSahara @croyle @ExplorElements @LostWorldExp @OverlandSphere @africaoverland

Crunched a bunch of data on #overlanding @LostWorldExp @OverlandSphere @africaoverland @croyle @ExplorElements pic.twitter.com/TZ81kXTWSd

— MJ (@mjfreshyfresh) October 25, 2014

[mjfreshyfresh]

To Build A PiLight

I've got 6 of these individually addressable units hanging in a tree out front. The #IoT for outdoor consumer lighting. Sourced parts from all over, one of the best places was right near my place The light Bulb shop on Mississippi Street, Portland, OR.

After experimenting with various materials I landed on these...

Top disc - local - TAP Plastics

Mounting hardwar - local - The "Light Bulb Lady" aka Sunlan Lighting

LEDs & Socket - Superbrightleds.com

Globes - ebay lightsnglobes

Silicon - local - hardware

[mjfreshyfresh]

Clean Street Art

As most of you know I’ve got a thing for street art. A while back I wrote a piece on Medium where I encouraged folks to develop a “street art eye” to make urban living more interesting.

Over the past few years I've also posted 100% street art on Instagram which covers my travels around the world.

I’ve always been an outsider looking in however last month I decided to try my hand at power washing stencils. Which seems to be a thing.

Drawing inspiration from a favorite silver bracelet I picked up in Thailand I set about creating a stencil from a piece of plywood. Using a hole saw I cut numerous holes in the plywood. Slapped it down on my driveway and had my wife pressure wash the hell out of it. Viola!

[mjfreshyfresh]

PiLights: Interactive LED Art

One of the things that really struck me living in Austin, Texas was the outdoor lights. Restaurants and private residences invested heavily in decorative lighting so that you could enjoy the Texas nights. That experience coupled with my trips to Burningman were the inspiration for PiLights.

For me the first step from an idea to reality is a whiteboard sketch. Above you can see the first iteration. Basically I want to:

Hang 6 spheres from the tree in my front yard.

Use 12 volt LED technology for illumination

Use RaspberryPi to make it interactive.

I'm now several months into the project. As always it's taking longer and costing more than expected. Fortunately I am getting close. More to come later...

[mjfreshyfresh]

Portland Short List

Making a trip to Portland, Oregon? Wondering what to do? Here are few suggestions...

Spent a day exploring downtown. If it's a Wednesday be sure to hit the downtown farmers market or if it's a weekend don't miss Saturday Market. Two must-see items around town are Powell's and VooDoo Doughnuts although I'd never wait in a long line for Voodoo. You'll notice numerous food carts, Portland is quickly becoming the food-cart-capital-of-the-world, be sure to eat from the food carts!

While you're downtown work your way NW and swing through the Pearl District with it's chic galleries, eateries and lack of homeless-hipsters you don't even know you're in PDX. Keep walking NW and you'll soon hit NorthWest Trendy Third some decent shops with good people watching. Forest Park is also quite close.

Portland is known as the City of Roses, be sure to check out some of the rose gardens

Another thing Portland is known for is it's Strip Clubs there are more per capita than anywhere else in the United State due to they way that the Oregon Supreme Court interpreted the first amendment "nude dancing is protected as free speech". If you want to scratch that itch Mary's Club is centrally located and has some history. Be sure to tip.

As for drinks... my favorite beer is the C-Note IPA by Lompoc Brewing, they have several locations swing by one for pint. Best coffee is Stumptown which you can now find in New York. But honestly I've yet to find bad beer or bad coffee in Portland. (some of it is questionable but not outright bad)

As for food.. too many choices, my top pick is Toro Bravo again you can't really go wrong.

My absolute favorite district is the Mississippi + Williams area which happens to be my neighborhood! Walk down Mississippi St. Stop in at the shops, drink, eat, be merry!

Now that you've explored Portland for a day or two get out of Portland! Do a day trip to the coast and check out Oswald West State Park. Or head to the mountains! Mount Hood is easily accessible and beautiful. (Just under 2hrs to either the Mountains or Coast from Portland). If you do make it out to Mount Hood then why not swing through Hood River on your way home.

If you're looking to do some camping here are a few trip reports from my adventures:

Coastal Camping

Olallie Lake

Barlow Road

(Photo credit: Ian Sane)

[mjfreshyfresh]

Data Science is for Everyone

Little Bird has been hosting the Portland Data Science Meetup for the past few months. I’ve had the great pleasure of seeing presentations such as Homophily: exploiting patterns of similarity by Devin Gaffney and getting to network with many interesting folks.

It thrills me to hear about the various ways people are leveraging graph analysis to solve problems. Just last week I met:

Plant Physiologist - Predicting plant health via graph. If one plant is unhealthily what is the likelihood of spreading? How can we better understand health over large areas?

Telco Ops - Optimizing for various call types. Long personal calls. Short support calls.

PhD Graduate - Years of schooling with very deep knowledge, looking to put it into practice.

Independents - A marketer who leverages data driven approaches to drive better results for clients.

REI Clerk - Who is bootstrapping a startup to drive better gear recommendations.

That’s just 5 of many conversations I had last week. Come on out to the next Portland Data Science Meetup and plug into the network.

[mjfreshyfresh]

Outerbody experience with a Dropcam

I’ve really been digging my Dropcam, the unofficial API is easy, Dropsy has been a hoot and I’ve had fun moving it around the house.

Inspired by Jason Wilson’s Outerbody Labs in which you get to "See Yourself Seeing Your Self” I set about hacking one together myself.

Unfortunately the lag on the Dropcam made it difficult and I ended up getting nauseous. Regardless it was still interesting enough and something you can do with your Dropcam for kicks.

Items:

1 Dropcam

1 iPhone + Dropcam app

1 empty cereal box

1 scissors

1 roll of tape

Cut a contour in the top part of the cereal box for your nose + cheeks and forehead.

Point Dropcam in the center of the room.

Turn on Dropcam iPhone app, sideways large.

Put rolled up tape on back of iPhone and place in bottom of cereal box.

Put box on face, look at iPhone, see yourself walking around the room.

Get nauseous and stop.

[mjfreshyfresh]

A Circular view of 10 Wearable Tech Companies and their 20 Influencers

About a month ago I took the top 10 Wearable tech companies and collected their 'friends' on Twitter. As you know your 'friends' on Twitter are the accounts that you follow. Meaning you wish to receive a flow of information from these accounts. In essence these accounts have some level of influence.

By analyzing thousands of Twitter accounts I was able to determine which accounts were most influential in the Wearable tech space.

Fast-forward to earlier this week when my father turned me on to CIRCOS one of several fantastic dataviz software packages by the brilliant Martin Krzywinski. CIRCOS was originally developed for visualizing genomic data but has quickly been adopted by the broader scientific community.

Since I had the Wearable tech data laying around I decided to give CIRCOS a spin. Following the data format guidelines I was quickly able to produce a format for CIRCOS to consume. The picture you see is the end result.

At a glance you think "aww pretty picture" but upon further inspection you realize there is a lot of information being conveyed.

Gizmodo, Techcrunch, Goodhealth, Womenshealth Runnersworld are the most influential.

Bodymedia, Withings, Ourlark follow the most influencers.

Tech Publishers, Health Publishers, Digital Health are the three major influencer classifications.

NOTES:

UPPERCASE - Wearable Tech Twitter accounts (truncated)

lowercase - Top influential Twitter accounts (truncated)

[mjfreshyfresh]

My First Time: A 3D Printed Gift

You always remember your "first time". Especially when it's experiencing a revolutionary new technology. I clearly remember getting our first VCR and being able to watch movies at home! Or getting my first iPhone and experiencing the magic of Apple.

Just the other week I had another first: My good friend Chris gave me a screw driver... THAT HE PRINTED AT HOME. Yes, my first 3D printing experience. Seeing how far we've come with the VCR makes me very excited about the future of 3D printing. BTW: You can print your own, it's on Thingiverse

[mjfreshyfresh]

Dropcam Deployed: First Impressions

Meet Dropsy the Dropcam. The latest #IoT addition to the HeckleHouse. I wanted to see what goes on "after dark" on our quiet little street, so I originally ordered a cheap camera and tried hacking things together myself. After buying several adapters and investing numerous hours I gave up on "rolling my own" camera system and started to look for a better solution. A good friend recommended Dropcam, after reviewing the website I was sold and promptly ordered one.

Two days after placing my order it arrived! Ten minutes later I had it deployed and was able to watch live video on my iPhone. Initial impressions are:

Great packaging and instructions. (very Applesque)

Desktop and mobile apps are simple and elegant.

Seven day "cloud recording" makes data management easy.

$300 is pricey but the end-to-end video solution is very compelling.

Really wish they had an outdoor camera.

The "Activity Recognition" looks very promising.

Now what really tickled my inner-geek was the fact five minutes after deploying the camera I was interfacing with it via the unofficial API by Nolan Brown. A few lines of code enabled me to grab photos from the camera and save them on my machine for further processing with tools such as OpenCV.

Needless to say I'm excited about this new #IoT addition and look forward to hacking on it this Winter.