Analysis APIs

Integrate with our Machine Learning powered services


Our APIs

All our analysis services are RESTful APIs powered by Machine Learning.

Text analysis

Our text analysis API offers sentiment analysis, subjectivity analysis, language detection, POS (parts of speech) detection, aspect analysis and aspect based sentiment analysis.

Image analysis

Our image analysis API offers common object detection and custom object/logo detection.

Text analysis

Sentiment & subjectivity analysis

Find the sentiment (positive, negative or neutral) and subjectivity (subjective or objective) of any text. We'll also return the confidence score of each result so you can decide if it meets your criteria.

Language detection

Find out which language text is written in. This service returns the ISO code for the language of the submitted text.

POS (parts of speech) detection

Find out the parts of speech of the submitted text. We'll tokenise the text you submit and advise the POS (e.g. NN - noun, NNS - plural noun etc.) of each token. We use the industry standard Penn Treebank II tag set. This service also splits the input text by sentence.

For example, submitted text of: "I love my cat."

Would result in a response containing:

  • "I": PRP (pronoun personal)
  • "love": VBP (verb, non-3rd person singlular present)
  • "my": PRP$ (pronoun, possessive)
  • "cat": NN (noun, singluar or mass)
  • ".": . (full stop)

Aspect detection

Find out which aspects are metioned where in the submitted text. When you submit text, you must specify the aspect set you wish to use for the detection. Current supported aspect sets include hotels, restaurants and airlines; a full list can be found in the documentation (available on request).

For example, submitted text of: "The hotel was only a five minute walk from the terminal. I enjoyed the lobster bisque in the restaurant which was quiet and beautifully decorated. The room was a little cramped and there were no tea/coffee making facilities."

Would result in a response containing:

  • Location: The hotel was only a five minute walk from the terminal
  • Food & Drink: I enjoyed the lobster bisque in the restaurant which was quiet and beautifully decorated
  • Decoration: I enjoyed the lobster bisque in the restaurant which was quiet and beautifully decorated
  • Accommodation: The room was a little cramped and there were no tea/coffee making facilities
  • Amenities: The room was a little cramped and there were no tea/coffee making facilities

Aspect based sentiment analysis

Find out which aspects are mentioned in the submitted text, and the sentiment associated with that aspect. The same aspect sets are available as for the aspect detection above.

For example, submitted text of: "The receptionist was extremely helpful. The shower and bathroom were very clean. The meal I had in the restaurant could have been much better."

Would result in a response containing:

  • Staff: Positive
  • Amenities: Positive
  • Food & Drink: Negative

Image analysis

Object detection

Will identify 90 common objects in the submitted image. A full list of supported objects is available in the documentation (available on request). The service can return an annotated version of the submitted image with bounding boxes around the detected objects (as shown below), a list of the objects detected, or a combination of both.

Custom object/logo detection

Similar to the object detection above but we can offer custom object or logo detection. If there is a particular type of object or brand logo you need to detect, let us know and we'll train our service to recognise it. The service returns the same output as the object detection above. An example of our service detecting the KFC logo and colonel is below.

Pricing

Our services are priced on a per call basis; there are no monthly fees or contracts so you only get charged for the calls you make. To discuss your requirements and get a personalised quote, please contact us.