This stock sentiment report will be analysing Qualcomm’s stock for the week commencing 2nd August 2021, using StockGeist.ai stock news sentiment API data. The report has been written by Finance Focused. The StockGeist platform determines the market sentiment indicators towards a stock in real-time by applying natural language processing to the latest news updates and social media posts. All times given are in GMT+0.
Qualcomm ($QCOM) Company Overview
Qualcomm ($QCOM) is an American multinational company that operates in the technology sector and more specifically the semiconductor industry. Founded in 1985 and headquartered in San Diego, California it develops wireless communication technology including, 4G and 5G, which connect smartphones to cellular towers and networks. Qualcomm is a component of the S&P 500 and NASDAQ 100 stock indices.
Historical analysis – 2020 and Jan-Jul 2021:

To start 2020, Qualcomm opened at $89.05 and continued the sideways movement at the end of 2019 until it started to decline with the broader market in February 2020 due to the Coronavirus pandemic, bottoming out at $58.00. From there, Qualcomm stock began recovering in an upward trend, also with the broader market.
On July 30th 2020, the stock jumped over 10% to $107.19 after reporting stronger than expected results and resolved a licencing dispute with Huawei. The bullish upward trend then continued as earnings expectations were beaten once again on November 4th, with the stock rising over 11% to $145.41 after earnings per share came in at $1.45 against an expected $1.17. At the same time, the company provided guidance on earnings estimates for the fourth quarter of an increase to between $1.95 and $2.15, which is likely to have fuelled the final upside movement before the stock price topped at $167.94 in January 2021.
On February 3rd, a downtrend began despite a good earnings report. Qualcomm’s next earning’s report fell short of expectations due to industry-wide supply constraints. Up to this point, Qualcomm’s stock rose 83% year to date (ytd). Meaning that investor expectations were set incredibly high and the earnings fell short of that high bar due to industry-wide supply constraints. The ensuing fall in the stock down to $122.9 on March 8th can be seen as profit taking after the spectacular rally that occurred in 2020. Despite this fall, the stock was still up over 70% ytd. The stock then seemed to move in a sideways range between approximately $122 and $145 for the next several months.
Recent analysis – 29th Jul 2021 to 6th Aug:

On July 29th Qualcomm broke out of this range, as shown by some simple technical analysis with the resistance now turned support level at the $145 handle. This was due to a strong earnings report on Q2 2021 in which it doubled its earnings by improving its access to foundries whilst the global semiconductor shortages continues. Resultantly, the stock rose 6% to a high of $151.52 before settling down to around $146.28 on Aug 6th.
Methodology
The impact of incoming emotional and information messages on the price of Qualcomm was assessed by gathering sentiment data, proposing a model for detecting changes in the number of incoming messages and then analyse the changes in the stock price.
The incoming messages are first classified as either positive, neutral, or negative by StockGeist’s AI deep learning-based solutions, an easy concept to grasp for sentiment traders. This data can be accessed through StockGeist’s dashboard, or it can be retrieved by using their REST API. An advantage of the API over the dashboard is that it includes more metrics such as positivity index (pos_index) and a normalised moving average difference (ma_count_change). The positivity index value reflects the interplay between positive, neutral, negative, and total messages [1]. Also, the normalised moving average count is the moving average (MA) of the difference in the MAs values of the total message count normalised by MA values [1].
Initial areas of interest were found by applying a Gaussian filter on the incoming stacked emotional and information messages, generating a stacked area steam chart. Even though we are interested in high frequency components when we propose a model for the data, the resulting graph gives a deeper insight by smoothing the curve and making it easier to ignore small sharp peaks.
The proposed trading strategy using sentiment data involves 3 signals. An initial signal if the moving average is increasing, a supporting signal if the normalised moving average difference is above 3 and a confirmation signal if both the total emotional and information messages are twice the current moving average. Upon confirming that a spike in the total incoming messages has occurred, we assess the messages using the positivity index. If the positivity index is above 1, it would result in buying the current stock and vice versa.
Finally, changes in the stock price were only evaluated once a suitable model has been fitted to avoid data snooping bias. The changes were calculated using hourly close price % change.
Graphs






Accuracy Test
Using the steam charts, Fig 2, and Fig 4, we were presented with three points of interest, 29th July, 2nd Aug and 5th Aug. These were chosen due to the increased number of messages in the analysed time frame. The following will break down the events hour by hour where the signals described in the methodology section were triggered.
On July 29th a very high influx of messages was recorded. However, we will ignore this area in the analysis as it may be momentum from previous days as shown by the moving average lines in Fig 1 and 3. Also, this date is not in the time frame that is being analysed in the report, but it adds to the understanding of the current week’s messages.
Next, at 11am 2nd Aug, a first signal was passed due to an increasing MA. Later, a supporting signal at 2pm was triggered by the normalised moving average difference reaching 3.76. The confirmation signal only was reached at 4pm where the information messages reached 109 and emotional messages stood at 44 with the moving average at 11.5. Finally, an assessment of the sentiment using the positivity metric was made. At 2pm the positivity index was 1.78, implying a buy and at 4pm it decreased to 0.48, suggesting a sell.
A few days later, we reach our third point of interest. Between the times 10am to 12am on 5th Aug we saw an increasing moving average. At 1pm 2nd Aug we received a very strong supporting signal but the specified confirmation signal was not quite reached. The normalised moving average increased from -0.44, the hour before, to an exceptionally high 7.45. However, 77 information messages and only 12 emotional messages were received with the total moving average at 9.58. The assessment of the positive index, which had a value of 0.475, suggests a sell at 1pm.
After proposing a trading strategy using data from the platforms stock sentiment features, we evaluated the performance of these choices on the stock prices using Fig 6. The predictions based on the sentiment data were correct and the changes in prices were greater than the changes around the times that were analysed as highlighted in green and red in Table 1 below. The refinement of the confirmation signal is needed since the current one will not always trigger on time. This will result in the market reacting faster than the algorithm. The API also has time series data that uses 5-minute intervals which would help capture sudden changes.
Overall, more back testing is needed to support that this type of sentiment analysis could result in a profitable trading strategy. However, there is some evidence suggesting that the tone of news articles posted online affects the price of the chosen stock.
Date | O | H | L | C | Change | V |
1st 19:30 | 149.68 | 149.98 | 149.50 | 149.8 | +0.09% | 117.665k |
2nd 13:30 | 150.14 | 152.28 | 149.92 | 151.90 | +1.40% | 120.418K |
2nd 14:30 | 151.96 | 151.99 | 151.17 | 151.42 | -0.32% | 49.708k |
2nd 15:30 | 151.53 | 151.83 | 148.34 | 148.80 | -1.73% | 140.612k |
2nd 16:30 | 148.55 | 150.00 | 148.83 | 149.72 | +0.62% | 89.953k |
4th 19:30 | 148.14 | 148.47 | 148.03 | 148.32 | +0.12% | 46.519k |
5th 13:30 | 147.34 | 147.41 | 144.59 | 145.30 | -2.04% | 142.723k |
5th 14:30 | 145.29 | 146.26 | 145.11 | 146.15 | +0.58% | 62.069k |
Reflections
Using a moving average to test for an increase in web traffic is quite weak, fitting daily sinusoids would be a better approach which becomes clearer once you have removed high frequency components of the incoming messages as shown in the steam charts. There are peaks during middays and troughs during nights. However, this sinusoidal nature is very surprising as it shows that all the web traffic is coming from a single time zone. As Qualcomm is an international company, the graph should be smoother due to news articles being posted at different times in different countries. After inspecting all the articles from 29th July about Qualcomm that StockGeist has analysed, all articles were found to be from English sources. This should always be considered when using the sentiment data, as international articles could affect the market more or even faster than news from only English sources.
With the recent breakout of the trading range of the past few months, caused by the company’s latest earning’s report, this simple but potentially influential technical for the stock is bullish. From a fundamental perspective, 5G will be the driver for Qualcomm’s stock going forward. The company is a global leader in 5G with over 150 licence agreements already signed, its handset revenue has risen 57% year-on-year to $3.9 billion (largely thanks to the adoption of 5G and with the technology has seen strong adoption in key markets including China. After a rough and patient testing 2021 so far for the stock, there is potential for a new upward trend to emerge.
Disclaimer: This report and analysis of Qualcomm’s stock was performed to determine whether positive messages detected by StockGeist translated into an increase in stock price, and whether negative messages translated into a decrease in stock price in this article’s analysis. Please note that this analysis is not investment advice.