Category: IoT

This history began in 2011 when the category IBM Champions for Social Business was created. At that time there were only 50.  During the following years I was also nominated in other categories such as IBM Middleware, IBM Cloud and recently IBM Watson IoT.

It is a journey of many years, a lot of learning and also knowledge sharing.

You can now package and deploy AWS Lambda functions as a container image of up to 10 GB.

This makes it easy to build Lambda based applications using familiar container tooling, workflows, and dependencies. Just like functions packaged as ZIP archives, functions deployed as container images will benefit from AWS Lambda’s operational simplicity, automatic scaling with sub-second startup times, high availability, and native integrations with 140 AWS services.

More information here

A device that has a temperature sensor in an office might report temperature in degrees Fahrenheit or in degrees Celsius. It is not efficient to configure applications to be able to consume data in all these formats. Instead, the data needs to be collected, transformed, and normalized to create a single logical model so that an application can interact with the different devices in the same way.

The data management component of Platform Service includes a device twin feature and an asset twin feature. The device twin feature lets you take advantage of the collection, transformation, and normalization of different formats of device data into a single logical model. The asset twin feature lets you group different devices together to create a Thing, which is a higher value asset-based data structure. You can even group Things together to create new Things. An application can interact with the logical model, regardless of the data format that is used by the individual devices or Things.

For example, a group of devices that report temperature, humidity, and ambient light can be aggregated into a “Room” Thing to represent the comfort level in a specific office. A number of “Room” Things can be aggregated into a “Floor” Thing to represent all offices on a specific level, and a number of “Floor” Things can be aggregated into a “Building” Thing. By using a Thing abstraction, your application is decoupled from the specifics of how the devices are connected, the format in which the devices publish event data, and how the data is combined.

The explanation above i got from the knowledge center, just to  introduce about a problem i  got last week.

The problem:  When we create device types  we must define a Hardware and a Logical interface. When we create  the interfaces we are modeling the schema of a table were data will be saved. You can define “the fields and his type”. As usual i create my interfaces, devices and put the sensors to send data, but some sensors are not sending one of the variables. For example the data expected was { voltage : value , kwh: value) but the sensor send {voltage: value}.  The data is not saved on the database.

To solve the problem i create a ticket and the answer was to test if a key exists if not save it as a null value.

On the logical interface, for each key i need to put the following (example for kwh) : $exists($event.kwh) ? $event.kwh(null)

“ImportError: No module named tensorflow” but you know you installed it? Sometimes you can import packages from the console, but not from the Jupyter notebook? !pip install tensorflow sometimes just does not work?

Googling about similar issues made me realise people are suggesting the thing that won’t work most of the time.

If you are installing packages by running

!conda install tensorflow
# or if you want to use pip
!pip install tensorflow

you are using very fragile commands (if run in notebook) and that’s the reason packages you installed can’t be imported. It is not fine this time

Instead, use these commands:

import sys
!conda install --yes --prefix {sys.prefix} tensorflow
# or if you want to use pip
!{sys.executable} -m pip install tensorflow

and you will not have problems with damn ImportError again.

Found this information on this link https://modelpredict.com/importing-packages-in-jupyter-notebook/

Sharpening a knife is a skill that everyone who enjoys cooking needs to have. I have several stones for sharpening my knives. In addition to having the correct stones one of the most difficult things is to maintain the angle between the knife and the stone during the sharpening process. There are many videos and even YouTube channels on the subject. A video that caught my attention is below. It shows how to create and use an electronic device to maintain the sharpness angle.

Rules are key elements for an IoT solution. Using rules you can trigger an action or send an alert any kind of notification.
IBM update his platform and now you can create rules using a graphical interface.

The key part of a rule is the expression that will be evaluate to true or false.  The expression language are built using JSONata.
JSONata is a lightweight query and transformation language for JSON data.

The point is that IBM are using JSONata v1.4x. and the current version of JSONata is v1.6.5. The versions have lots of difference.
Yesterday i was writing a rule to get the hour for an event. If you are using v1.6.5 you can use the function $fromMilllis with picture.
$fromMillis(1510067557121, ‘[H01]’, ‘-0500’) => “10”
But Watson IoT are using v1.4x  and $fromMillis doesn’t have the picture implemented. The only way i found is to manipulate strings and i got the following expression:

$number($substring($fromMillis($state.datahora),11,1)) < 1 and $number($substring($fromMillis($state.datahora),12,1)) < 3 ? $number($substring($fromMillis($state.datahora),12,1))+24-3 : $number($substring($fromMillis($state.datahora),11,1))*10+ $number($substring($fromMillis($state.datahora),12,1))-3
$state.datahora returns an epoch timestamp.
This expression returns the hour on my timezone (-3). If anyone have a more elegant solution post on comments. I know my solution can’t be the better one, but it works.