Trading Indicator API Beta


API Example

For a full fledged demo application that runs on our API, please visit: TechSentiment.com


Company Ticker Price Predicted return Financial Sentiment Consumer Sentiment Message Volume
Apple Inc.
AAPL N/A -16.004% ( 29.82 ) 11.912% ( 15591 ) 559108
Google Inc.
GOOG 900.36 N/A 36.833% ( 48.61 ) -11.356% ( -227 ) 237263

Last update: June 18, 2013, 8:15 p.m. (EST)



Google vs Apple Sentiment Breakdown

AAPL message volume vs. trading volume

GOOG message volume vs. trading volume



The data explained


The Trading Indicator API gives you real-time insights into human emotion in financial markets. Sentiment scores are given via a JSON-structured data feed that contains several elements that can be used in asset pricing, which are explained below. Also take a look at our demo application we've build with our API.


Sentiment Feature JSON object
Financial Sentiment Score
fin_sntmnt_score
The financial sentiment score is the aggregated score of all messages surrounding a specific stock that contain financial sentiment during the last 24 hours.

A high score indicates that financial sentiment is bullish, while a low (negative) score indicates that financial sentiment is bearish.
Average Financial Sentiment Score
fin_sntmnt_avg_score
The average daily financial sentiment score caculated over the past six months. This score can be used as a benchmark for "normal" financial sentiment levels surrounding a specific stock or asset.
Financial Sentiment Momentum
fin_sntmnt_momentum
The difference between the financial sentiment score over the past 24 hours when compared to normal levels (long term average). A high percentage indicates bullish sentiment divergence, while a low percentage indicates bearish sentiment divergence.
Consumer Sentiment Score
comm_sntmnt_score
The consumer sentiment score is the aggregated score of all messages surrounding a specific company that contain consumer sentiment during the last 24 hours.
A high score indicates that consumer sentiment is bullish, while a low (negative) score indicates that consumer sentiment is bearish.
Average Consumer Sentiment Score
comm_sntmnt_avg_score
The average daily consumer sentiment score caculated over the past six months. This score can be used as a benchmark for "normal" consumer sentiment levels surrounding a specific company.
Consumer Sentiment Momentum
comm_sntmnt_momentum
The difference between consumer sentiment over the past 24 hours when compared to normal levels (long term average). A high percentage indicates positive consumer sentiment divergence, while a negative percentage indicates a negative divergence in consumer sentiment.
Financial Message Volume
fin_volume
The total number of messages that contained financial sentiment during the last 24 hours.
Average Financial Message Volume
fin_avg_volume
The average daily number of messages that contained financial sentiment during the last six months.
Financial Message Volume Momentum
fin_volume_momentum
The difference between the financial message volume over the last 24 hours compared to the long-term average. A high percentage indicates that there's a lot of financial buzz surrounding a company or stock, while a low percentage indicates that there's relatively little buzz.
Consumer Message Volume
comm_volume
The total number of messages that contained consumer sentiment during the last 24 hours. Consumer message volume is usually a lot higher than financial message volume, as there are more consumers talking about companies than investors talking about stocks.
Average Consumer Message Volume
comm_avg_volume
The average daily number of messages that contained consumer sentiment during the last six months.
Consumer Message Volume Momentum
comm_volume_momentum
The difference between the consumer message volume over the last 24 hours compared to the long-term average. A high percentage indicates that there's a lot of consumer buzz surrounding a company or stock, while a low percentage indicates that there's relatively little buzz.
Total Message Volume
total_volume
The total number of messages surrounding a stock during the last 24 hours.
Average Total Message Volume
total_avg_volume
The average number of daily messages surrounding a stock during the last six months.
Predicted Return
predicted_return
The predicted return for the upcoming hour, based on the above sentiment features and actual stock price moevement during the last five hours. The returns are given as a percentage, and are only availabie during NASDAQ opening hours. For more information how the predictions are constructed, please see our Methodology section.
Confidence Interval
prediction_confidence
Each prediction is accompanied by a confidence interval. The confidence interval indicates how ceirtain our algorithms are that the predicted return is correct. If the confidence level is 55 percent, 55 out of a hundred times the prediction should be correct.
  1. {
  2.   "data": [{
  3.       "stock_id": 163,
  4.       "ticker": "AAPL",
  5.       "name": "Apple Inc.",
  6.       "fin_sntmnt_score": 0.58378776,
  7.       "fin_sntmnt_avg_score": 68.7788001631164,
  8.       "fin_sntmnt_momentum": -0.38382559433322,
  9.       "comm_sntmnt_score": 11239.1272482,
  10.       "comm_sntmnt_avg_score": 21010.6054282,
  11.       "comm_sntmnt_momentum": -0.458908044062,
  12.       "fin_volume": 20.0,
  13.       "fin_avg_volume": 789.545454545455,
  14.       "fin_volume_momentum": -0.37167961606003,
  15.       "comm_volume": 11371.0,
  16.       "comm_avg_volume": 247246.947368421,
  17.       "comm_volume_momentum": -0.8438096879189,
  18.       "total_volume": 11391.0,
  19.       "total_avg_volume": 248036.49282296645,
  20.       "predicted_return": 2.950315372003879E-5,
  21.       "prediction_confidence": 0.52759705647226
  22.   }]
  23. }


Historic cases for the Dutch AEX Index




Case I - SNS Reaal (SR.AS)





A

B

C


D




Case II - Royal KPN (KPN.AS)





A

B




Use case III - Royal Dutch Shell (RDSA.AS)


A



Our methodology explained



Multiple studies over the past years have shown that Twitter sentiment can be used as an indicator for future stock price movement. SNTMNT has developed such an indicator: the first API in the world that gives hourly and/or daily buy & sell signals for all S&P 500 stocks based on online buzz on Twitter. These predictions take away extra noise in asset pricing, and give investors an additional trading indicator on top of fundamental analysis and/or technical analysis.

Example messages that show how our algorithms classify messages positive (bullish), neutral or negative (bearish):



Positive (bullish) messages
Neutral messages
Negative (bearish) messages


The methodology that we apply is best explained in three steps:

Step 1 - Sentiment Analysis
First we use NLP algorithms to classify whether the Twitter feed surrounding each individual S&P500 fund is positive or negative. We do this by looking both at financial sentiment and brand sentiment surrounding a specific company.

Step 2 - Reach and Authority
For each tweet, we look at its authors reach on Twitter, by using indicators like Klout and PeerIndex. This ensures that we’re not just looking at news creators, but more so at news recipients. To determine authority we also take into account the author’s field of expertise (topics).

Step 3 - Machine Learning
We use machine learning methods, including robust non-linear models, to look at correlations between Twitter sentiment, author authority and the S&P 500. The predictions that result from these correlations can have an accuracy as high as 60 percent, with an average of 54 percent. Each prediction is accompanied by a confidence interval that reflects the reliability of the prediction.


SNTMNT Trading Indicator API Methodology

Why are we different

We are inspired by the work of professor Johan Bollen, who found correlations between the Dow Jones Industrial Average and Twitter sentiment. There is a big difference though between his models and ours. Where Bollen focuses on a very macro level of Twitter sentiment (general mood), we believe that more value can be extracted by focussing on a more micro level. Our predictions are based on very specific fund related Twitter sentiment through stock tickers, in the way StockTwits does, and brand related sentiment (name, products, CEO).



Historical time series for backtesting



It is possible to receive historical sentiment data that can be used to backtest our sentiment analysis on your predictive models. Just drop us a line, and we'll discuss how we can help you best.






Our Social sentiment indicator is available for Forex, S&P500, NYSE Euronext and Commodity markets. Interested to test it's performance on your platform?


Contact us about an API-key