Welcome to the new SCiO Developer Blog!

My name is Hagai Heshes and I’m responsible for the developer community at Consumer Physics. The team and I will be blogging here to keep you informed about the SDK, provide behind the scenes insights into our own development process, share technical information and, in general, keep you up to date.

First, a quick update on our product names – based on feedback, we renamed the developer kits to make things more clear:

• “Application Development Kit (ADK)” is now “Software Development Kit (SDK)”
• “Researchers Kit” is now “Researcher SDK”

The SDKs enable you to develop molecular sensing models and/or mobile apps.

• Molecular sensing models are chemometric models that interpret SCiO scan results, such as the quality of olive oil, sweetness level in green apples and many other possibilities. As developers, you can create basic models by using the automatic processes supported by the SDK, or create more advanced models by analyzing the scan results (the reflected spectra) and manually refining them using the software tools provided within the SDK. The Researcher SDK also enables you to download the spectra results to your computer for analysis with your own tools.
• Mobile apps for iOS and Android can be developed using the SDK mobile libraries that enable communicating with the SCiO molecular sensor and the database of molecular sensing models. These apps can use models developed by you, models developed by Consumer Physics, or models developed by other developers who shared them with the SCiO community.

You can choose to develop only molecular sensing models, or only mobile apps based on existing models, or both.

 

SCiO SDK Workflow

The question we hear often from developers is “what’s the workflow?”
That depends on whether you intend to develop molecular sensing models, mobile apps or both. The flow below outlines the steps for developing both. Steps 1 and 2 are for developing the models. If you are only interested in developing apps, you’ll skip directly to step 3.

 

Step 1: Collect Data

In this step, you gather the data for your experiment. This includes defining what you want to test and the attributes that are of interest to you (e.g. sugar level, country of origin). This step also includes scanning your samples and filling in the corresponding attribute values to your scan results. For example, if you’re interested in classifying oil types (e.g. olive oil, canola oil, etc.), you first define one or more attributes such as “oil type” and “oil quality”, scan your oil samples and for each scan result you will provide type and quality.

 

Step 2: Develop Molecular Sensing Models

Once you have the data collections from the previous step, you will analyze this data using a GUI and apply a variety of algorithms, perspectives and other tools that are part of the SDK in order to create a molecular sensing model that can be used to interpret scans. So, using the oils example again – you will analyze the oil records you created in step 1 until you are confident that you have a model that can interpret oils the way you intended it to work.

Nissan, Roni and Amit analyzing spectra at Consumer Physics HQ, while creating a molecular sensing model for milk

 

Step 3: Use Models

Once a model is available for use, you can develop apps for iOS and Android using our mobile libraries. These apps will communicate with the SCiO molecular sensor to get scan results and can be based on one or multiple molecular sensing models. The models may be the ones developed by you, as described in the previous steps, or models developed by Consumer Physics or other developers.

 

That’s it for now – a high level introduction to the SCiO SDK workflow. Stay tuned for more details in future posts.

One last thing. We are currently building the SCiO developer forum that will go live soon and we would love to get your input on the topics you want included. Send us the hot topics you expect to see and want to discuss tohack@consumerphysics.com.

Happy New Year from the entire Consumer Physics team!
Hagai