Tick Data Analysis: Calculating Multi‑Short Positions with Python

The article treats tick data as a battlefield where each tick represents a trade action, emphasizing the importance of understanding both sides (long and short) in real‑time market analysis.

It introduces the basic fields of a tick (time, price, volume, position change, and nature) and outlines a four‑step method to calculate multi‑short and multi‑long volumes:

Identify and record non‑directional open/close actions (e.g., opening, closing, double opening, double closing, turnover, unknown).

Determine price direction (up, down, unchanged) based on comparisons between the latest price and the best bid/ask.

Combine the results of steps 1 and 2 to assign directional open/close labels (e.g., long open, short close, unknown).

Calculate the actual open/close volumes and visualize the results.

Preparation includes using PyCharm, importing pandas , matplotlib , and talib , and loading the tick CSV file into a DataFrame. The following code shows the initial data loading and cleaning:

<code>import pandas as pd
import matplotlib.pyplot as plt
import talib
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
count = 6000
RB_data_csv = pd.read_csv('RB1910.csv', encoding='UTF-8')
RB_data = pd.DataFrame(RB_data_csv)
# Data cleaning
time = []
check_close = []
position = []
volume = []
close = []
sellshort_1 = []
buy_1 = []
for i in range(len(RB_data['时间'])):
    time.append(RB_data['时间'][i])
    position.append(RB_data['持仓量'][i])
    volume.append(RB_data['现手'][i])
    close.append(RB_data['最新价'][i])
    sellshort_1.append(RB_data['卖一'][i])
    buy_1.append(RB_data['买一'][i])
</code>

Next, the script computes the per‑tick hand volume ( Now_hand ) and position change ( Entrepot ) by differencing successive ticks:

<code>Now_hand = []
Entrepot = []
for i in range(len(volume)):
    if i == 0:
        Now_hand.append(0)
        Entrepot.append(0)
    else:
        Now_hand.append(volume[i] - volume[i-1])
        Entrepot.append(position[i] - position[i-1])
</code>

Using these arrays, the script classifies each tick’s non‑directional type ( trade_type ) with conditional logic, then determines price direction ( trade_direction ) and finally assigns a directional type ( trade_netrue ) such as "多开" (long open), "空平" (short close), "双开" (double open), etc.

<code># Example of directional classification
if trade_type[i] == '换手' and trade_direction[i] == '向上':
    trade_netrue.append('多换')
elif trade_type[i] == '开仓' and trade_direction[i] == '向上':
    trade_netrue.append('多开')
# ... (other conditions omitted for brevity)
</code>

Finally, the article shows how to translate the directional labels into actual volume series for long buys, short sells, covers, and unknown actions, and presents visualizations of the aggregated multi‑short/long volumes over the trading day.

Overall, the tutorial provides a complete pipeline—from raw tick CSV to cleaned DataFrame, through stepwise classification, to volume calculation and plotting—enabling quantitative analysts to extract market micro‑structure signals from high‑frequency futures data.