加密貨幣 - 情緒指標

在沒有時間盯盤以及頻繁做短線進出情況下，較適合的方式不外乎定期定額，若可以搭配情緒指標，在市場情緒低迷時，可以以較低成本的金額投資商品，並且在市場情高昂時高歌離席。

常見的加密貨幣情緒指標有恐慌指數(Greed and Fear Index) 與資金費率(Funding Rate)。這裡介紹另一種觀察市場情緒的指標：應用程式排行。

應用程式排行

原理

加密貨幣交易所發行在行動裝置 iOS 以及 Android 作業系統的應用程式下載量（使用量）可依照其排行而定。若排行越前面，其使用量越高，代表越多用戶參與這個市場。反之排名越後，使用量越低，越少用戶參與市場。
應用程式排行可以充分反應市場情緒，參與市場者越多，象徵市場情緒升溫，隨時有高點反轉的可能，是個警示訊號。反之參與市場者越少，象徵市場情緒過冷，可以適時搭配其他指標進場布局。
這裡使用 Binance 台灣在 iOS 上的排行為例，以 Python 抓取數據並分析加密市場的情緒熱度。

資料來源

以網站 Sensor Tower 每日公佈的應用程式排行為資料來源，使用 Python 抓取 Binance 台灣在財經類別的排行，分析用戶參與市場的熱度。

數據處理

使用 Python 的 requests 函式庫，爬取網站資料：

url = 'https://app.sensortower.com/api/ios/category/category_history?app_ids%5B%5D=1436799971&categories%5B%5D=6015&categories%5B%5D=0&categories%5B%5D=36&chart_type_ids%5B%5D=topfreeapplications&chart_type_ids%5B%5D=topfreeipadapplications&chart_type_ids%5B%5D=toppaidapplications&countries%5B%5D=TW'

response = requests.get(
    url=url
)

轉為 DataFrame 格式：

df = pd.DataFrame(response.json()['1436799971']['TW']['6015']['topfreeapplications']['graphData'], columns=['timestamp', 'rank', 'status'])
df.set_index('timestamp', inplace=True)
df.index = pd.to_datetime(df.index*1000000000, utc=True)
df = df.drop(['status'], axis=1)

我們有興趣的資料欄位需透過 1436799971， TW ， 6015， topfreeapplications
以及graphData 索引而得，並且捨去不需要的欄位。

df呈現的資料如下：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

分析

繪製排行歷史

df.plot(figsize=[20, 10])

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

由於網站限制存取 2 個月的資料，因此呈現效果有限。剛好手邊有個比較完整的歷史數據如下：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

在接近 2024 年有一天抓到了排名 500 的資料，推測是當天沒有排名或是沒有抓取到，因此移除此筆並沿用前一日的排名：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

與 BTC 走勢一起繪製：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

將排名曲線以 20 日平均繪製，去除雜訊：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

可以大致看出，排名曲線與比特幣走勢呈現反向。

我們也可以縮進 (zoom in) 至較短的時間區段，觀察走勢，較不易受全局的龐大尺度影響。

首先觀察 2017 年底至 2018 年初，此段歷經暴漲與暴跌。

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

可以看到 2017 年底的漲幅，伴隨著使用量上升，象徵用戶積極參與市場。從 2017 年底至 2018 年初，使用量一直維持著名列前茅，而後得跌勢也可以發現使用量大量下滑。

2020 年三月至十月，比特幣經歷了暴跌與大幅成長，兩者的走勢出現反向的趨勢。

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

2020 年九月至 2022 年 6 月，比特幣連續經歷了兩次 ATH 與 Luna 暴雷事件，兩者的走勢出現反向的趨勢。

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

Coinbase 數據

這裡只分析了 Binance 台灣的 iOS 使用情形，同時參考 Coinbase 於美國的使用情形，是否有同樣的走勢：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

將排名曲線以 20 日平均繪製，去除雜訊：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

直接觀察 2020 年九月至 2022 年 6 月：

Image Not Showing Possible Reasons

The image was uploaded to a note which you don't have access to
The note which the image was originally uploaded to has been deleted

可以看出 Coinbase 於美國的使用情形與 Binance 台灣的使用情形相似。

結論

應用程式排行雖無法指出可以精準抄底的時機，但對於逃頂有清楚的指示，且台灣與美國用戶的使用情形相似。

若能再納入 Upbit 於韓國的排行合併較完整的情緒指標。

程式碼連結

首先下載完整的 Binance iOS 應用程式排行：
TW Binance iOS
再至 Colab 連結：
Sentiment analyze
於左方檔案列表新增資料夾，可自行命名，此處以 data 為範例，將以上的應用程式排行檔案移至新增的資料夾。